Master Thesis - RUN: Página principal · 2019. 6. 26. · Master Thesis The effect of CSR...

1

Master Thesis

The effect of CSR performance on firm value across global

emerging markets

Erik Nielsen

32344

M.Sc. International Finance

December 2018

Supervisor: Melissa Prado

2

Abstract

Drawing on institutional and transaction cost theory and the resource-based view of CSR, I

posit that CSR performance across global emerging market companies is positively related to

firm valuation. Using an unbalanced panel data approach for a sample with 3,800 firm-year

observations representing 657 individual firms from 20 different countries that are classified as

emerging markets according to the MSCI EM index during 2010-2016, I find that CSR

performance, proxied by the average of the environmental and social pillar scores of the

Thomson Reuters EIKON ESG rating database, positively relates to firm valuation, proxied by

one-year ahead Tobin’s q (TOBQ). Specifically, a one-standard-deviation increase in

normalised CSR performance is – on average – associated with a 0.042-point increase in TOBQ.

Compared to the mean value of 1.661 for TOBQ across the sample, this increase constitutes an

economically significant share of around 2.5% of that value. This value-enhancing effect of

CSR is driven by companies in Asia, while it is absent for companies located in EMEA and

more pronounced for companies of the Americas. Additional analyses further reveal that while

overall Thomson Reuters corporate governance score performance is positively related to firm

valuation, the way in which these scores are constructed seems to fail to reflect important

differences in the governance environment of emerging market companies compared to their

developed market counterparts. Moreover, the number of sell-side analysts covering the stock

of these companies is (next to CSR) positively related to firm valuation but has a mitigating

effect on the positive relation between CSR performance and firm valuation.

Key words: Institutional theory, Transaction costs, Emerging Markets, CSR, Firm Value

Acknowledgements: I want to thank Prof. Rob Bauer (Maastricht University) and Associate

Prof. Melissa Prado (Nova School of Business and Economics) for accompanying me through

the past semester. I am grateful for all your inputs and your dedicated time. You found a great

balance between giving me enough space to work independently and providing me with

guidance if needed. It was a pleasure to work with and get to know you.

3

Table of Contents

1. Introduction .................................................................................................................................................... 5

2. Literature Review ......................................................................................................................................... 11

2.1 CSR in the context of SRI and ESG .................................................................................................... 11

2.2 The long-standing debate on CSP and CFP ......................................................................................... 13

2.2.1 CSR and stock performance ........................................................................................................ 13

2.2.2 CSR and firm risk characteristics ............................................................................................... 14

2.2.3 Advancements in ESG data quality/availability .......................................................................... 16

2.3 CSR in the societal, organizational, and institutional context of emerging markets ............................ 17

2.3.1 Cross-country differences in CSR activity .................................................................................. 17

2.3.2 Institutional context and commonalities in emerging markets .................................................... 18

2.3.3 Strategic responses of emerging market firms to overcome institutional voids .......................... 19

2.3.4 Societal and organizational trends driving CSR sensitivity across global emerging markets .... 20

2.4 The CSP and CFP debate in the institutional context of emerging markets ........................................ 21

2.4.1 The effect of CSR on firm value across global emerging markets .............................................. 21

2.4.2 The role of firm-level governance in emerging markets ............................................................. 22

2.4.3 The role of sell-side analysts on CSR and firm value across global emerging markets ............. 23

2.4.4 The role of institutional ownership in global emerging markets ................................................ 25

3. Research Design ........................................................................................................................................... 27

3.1 Sample construction ............................................................................................................................ 27

3.1.1 The sample .................................................................................................................................. 27

3.1.2 Measuring CSR performance ...................................................................................................... 30

3.1.3 Measuring firm value .................................................................................................................. 31

3.1.4 Control variables ........................................................................................................................ 32

3.2 Methodology ........................................................................................................................................ 33

3.2.1 Panel data analysis ..................................................................................................................... 33

3.2.2 Fixed effects OLS model ............................................................................................................. 34

3.2.3 Generalized least square regression (EGLS) .............................................................................. 35

4. Results .......................................................................................................................................................... 37

4.1 Descriptive Statistics ........................................................................................................................... 37

4.2 Inferential statistics .............................................................................................................................. 42

4.2.1 The effect of CSR performance on firm value across global emerging markets ......................... 42

4.2.2 The role of firm-level governance and ESG rating categories.................................................... 46

4.2.3 The role of sell-side analysts on CSR and firm value ................................................................. 49

5. Discussion & Limitations ............................................................................................................................. 51

6. Conclusion .................................................................................................................................................... 58

References ........................................................................................................................................................ 62

Appendix .......................................................................................................................................................... 68

4

List of Tables

Table 1: Construction of the unbalanced data panel ............................................................................................. 28

Table 2: Sample distribution by country, region, and industry group ................................................................... 30

Table 3: Descriptive statistics for the full and reduced sample ............................................................................. 38

Table 4: TOBQ and CSR by country and region .................................................................................................. 41

Table 5: Baseline and regional TOBQ regressions ............................................................................................... 43

Table 6: Year-by-year baseline TOBQ regressions ............................................................................................... 45

Table 7: Firm-level governance and ESG rating categories TOBQ regressions ................................................... 48

Table 8: Analyst TOBQ regressions ..................................................................................................................... 50

List of Figures

Figure 1: Model of Relationships .......................................................................................................................... 26

5

1. Introduction

Global institutional investors acting in the long-term interests of their beneficiaries increasingly

incorporate environmental, social, and governance (ESG) aspects in their investment analyses

and portfolio selection processes, next to traditional economic factors. Corporate social

responsibility (CSR) closely relates to the environmental and social non-financial performance

dimensions and is often defined as a mechanism to address externalities that firms generate in

the process of pursuing profit maximization that are not internalized by shareholders (Lian and

Renneboog, 2017). To incorporate ESG aspects into their investment decisions, portfolio

managers often rely on information intermediaries such as ESG rating agencies or analyst

reports. According to Morgan Stanley Capital International (MSCI) ESG, for example, 31 of

the top 50 institutional money managers worldwide use their research to integrate CSR factors

into their investment decisions (Kim, Li, and Li, 2014).

From a firm perspective, the central theme academics inspect is the relationship between

corporate social performance (CSP) and corporate financial performance (CFP). The initial

debate revolves around two contradicting views of the firm and the function of CSR, the

resource-based view and the negative view. The resource-based view states that CSR

investments create necessary resources and stakeholder support (e.g. Jones, 1995), that translate

into sustainable long-term competitive advantage and thus higher financial performance

through more favourable return and/or risk profiles (e.g. Waddock and Graves (1997); Russo

and Fouts, 1997; Deng, Kang, and Low, 2010). On the other hand, the negative view suggests

that companies should not internalize the negative externalities they exert on other stakeholders

than shareholders, as doing so conflicts the sole responsibility of firms to maximize shareholder

value (e.g. Pigou, 1920; Friedman, 1970).

The rationale of the resource-based view is that CSR investments create intangible assets

through various channels. Continuous stakeholder-management relationships induced by CSR

properly incentivizes mangers to focus on financial goals (Hill and Jones, 1992; Jones, 1995),

builds moral capital or goodwill amongst stakeholders in good times to draw on in worse times

(e.g. Godfrey, 2005), stipulates reporting information quality and disclosure (e.g. Gelb and

Strawser, 2001) and builds a reputation for quality, reliability, and trust in the market

(McWilliams and Siegel, 2001). Facilitated by the severe erosion of investor wealth and

numerous shutdowns of businesses during crises periods and their frequent occurrence, the

research on CSR activity shifted away from an initial return focussed debate towards the effect

6

of CSR on firm risk characteristics. Specifically, recent studies on the CSP CFP debate focus

on stock price crash risk (Kim et al., 2014; Diemont, Moore, and Soppe, 2016; Zhang, Xie, and

Xu, 2016; Lee, 2016; Utz, 2017) and span over stock market boom periods (1995–1999 and

2003–2007) and/or burst periods (2000–2002 and 2008–2009). At the same time, studies point

out that institutional differences between countries and regions drive the observed differences

in the extent, content, and communication intensity of CSR between companies (e.g. Liang and

Renneboog, 2017).

Institutions, or specialized intermediaries in an economy emerge in response to mitigate the

various transaction costs associated with market failures in the economic exchange process

arising from negotiating, monitoring, and enforcement costs (Coase, 1937; Williamson, 1975,

1985; Jones and Hill, 1988). Formal market-supporting institutions, such as capital markets,

regulatory systems, and contract enforcement mechanisms across global emerging markets are

generally weaker compared to developed markets (e.g. Meyer, Estrin, Bhaumik, and Peng,

2009). This results in business being guided rather by informal institutions such as business

groups, family connections and government contacts, which often creates governance concerns

via family- or government-controlled companies and ultimately hinders the proper protection

of minority shareholder rights (Young, Peng, Ahlstrom, Bruton, and Jiang, 2008). That is why

investors taking a minority stake in emerging market companies endure higher transaction and

monitoring costs of their investments compared to equivalent investments in developed markets

(e.g. Claessens, Djankov, and Lang, 2000; Faccio, Lang, and Young, 2001).

Khanna and Palepu (1997, 2011) argue that in emerging market companies must develop

strategic responses to overcome undue transaction costs and restricted access to resources

caused by the absence of market-supporting institutions. Strategic responses could be to build

a reputation of treating minority shareholders fairly (Gomes, 2000), strategic alliances with

foreign multinationals from countries with strong institutions (Siegel, 2009), geographical

clustering to create local business environments (Karna, Täube, and Sonderegger, 2013),

expanding the business abroad to access more efficient and munificent foreign markets (Luo

and Tung, 2007), or signalling commitment to fair practices by voluntary cross-listings on

exchanges with strong monitoring standards (Young et al., 2008). The very recent study of El

Ghoul, Guedhami, and Kim (2017) finds that CSR commitment is another strategic response to

reduce transaction costs and to tap additional resources. Specifically, they show empirically

that CSR performance is more positively related to firm valuation in countries with lower

market-supporting institutions.

7

While CSR awareness has been traditionally attributed to be confined to developed

societies/economies in the academic landscape (e.g. Arya and Zhang, 2009), momentum is

pronouncing the relevance of CSR across emerging markets. A recent CEO study of Lacy and

Hayward (2011) on the topic of sustainability in partnership with United Nations Global

Compact (UNGC) shows that more CEOs of emerging market companies describe

sustainability as important to their company’s future success compared to developed market

companies. The McKinsey Global Institute (2010) further argues that the demand for growth

capital across global emerging markets is expected to keep surging for the foreseeable future,

so emerging market companies keep on seeking investments from ESG sensitive investors

abroad. Moreover, increasingly international supply chains and business ties between emerging

and developed market companies draw attention from Western societies to the responsibility of

emerging market suppliers of Western companies (e.g. Bogdanich, 2008; Perez-Batres, Miller,

and Pisani, 2010). Moreover, the emergence of emerging market multinationals such as

Embraer, Tata, Alibaba and Haier gain considerable world-wide attention and increasingly fall

under the same CSR scrutiny as their Western counterparts.

Combining the recent finding of El Ghoul et al. (2017) that CSR performance is more positively

related to firm valuation in countries with lower market-supporting institutions with the fact

that global emerging markets – while cross-country differences exists – are characterized by

weaker institutions and the ongoing societal and organizational shift towards more CSR

sensitivity in these societies, theory strongly points towards a positive relationship between

CSR performance and firm value across global emerging market firms as a group. Yet, there is

no previous study examining the CSP and CFP debate in the context of global emerging markets

as a group. This is probably due to the failure to recognize institutional commonalities across

global emerging market companies, ESG data scarcity in emerging markets and the extremely

volatile capital markets in the last decade steering the focus on risk characteristics. This thesis

closes this research gap by examining the research question whether CSR performance across

global emerging market companies increases their firm valuation.

In the main empirical analysis, I use an unbalanced panel data approach for a sample of 3,800

firm-year observations representing 657 individual firms from 20 different countries that are

classified as emerging markets according to the MSCI EM index during 2010-2016. In line with

previous research (e.g. Kim et al., 2014; El Ghoul et al., 2017; Utz, 2017), CSR is defined as

the average of the Thomson Reuters EIKON environmental and social pillar scores. I use

Tobin’s q (TOBQ), which is the market value of the firm’s assets divided by the replacement

8

value of the firm’s assets as firm value proxy for financial performance, as it should capture

return and risk effects of CSR in aggregate (e.g. Gompers, Ishii, and Metrick, 2003; El Ghoul

et al., 2017). I follow previous studies in controlling for return on assets, firm size, leverage,

GDP of the respective head quarter country, and firm age. I find that CSR performance

positively relates to firm valuation. Specifically, a one-standard-deviation increase in

normalised CSR score performance is – on average – associated with a 0.042-point increase in

one-year ahead TOBQ. A robustness check of year-by-year regressions with ordinary least

squares (OLS) estimates confirms this result. However, this result is largely driven by the Asian

companies representing around 60% of the sample. Companies located in EMEA representing

20% of the sample do not show any value enhancing effect of CSR and companies located in

the Americas representing the remaining 20% of the sample show an especially strong value

enhancing effect of CSR.

Additionally, this thesis inspects the role of firm-level governance on firm valuation in

emerging markets. Institutional theory suggests that the need for effective internal corporate

governance is especially high in countries where institutional voids inhibit market oversight or

external governance mechanisms (Ding, Wi, Li, and Jia, 2010). However, the ESG rating scores

of Thomson Reuters are constructed via a consistent criteria list which is based on the

predominant corporate governance model of developed markets, where principal-agent (PA)

problems between owners and managers receive most attention (Jensen and Meckling, 1976).

In contrast, concentrated ownership of states and families in emerging markets is often the root

cause of expropriation of minority shareholders, which Young et al. (2008) term principal-

principal (PP) problems. A dissection of the overall governance score into its three categories

management, shareholders, and CSR strategy shows while the consolidated corporate

governance performance measure has a positive effect on firm value, there is some evidence

for a negative effect of the Shareholders score. This finding supports the view of Faccio et. al

(2001) that power towards shareholders vs. managers in emerging markets might not be as

positive as is believed according to developed markets theory.

Furthermore, this thesis inspects the role of sell-side analyst coverage on CSR and firm

valuation in an additional analysis. Using the same unbalanced panel approach for a reduced

sample with 3,504 firm-year observations reveals that analyst coverage is positively related to

firm valuation – next to CSR. Including an interaction term further shows that the number of

analysts has a mitigating effect on the positive relationship between CSR and firm value, while

the previously found positive effects stay significant. Taken together, these findings support the

9

view of Chung and Jo (1996) and Yu (2008), that equity analysts act as an external monitor and

help reduce transaction costs by reducing agency conflicts, disciplining managers and steering

investor attention towards important information. While analysts seem to provide some

additional social pressure on firms to reduce their irresponsible activities and signal trust in the

proper governance of the firm, they rather complement than substitute ESG rating agencies as

information intermediaries by focussing on financial reporting irregularities.

This thesis enhances the current academic debate and is highly relevant for practitioners such

as portfolio managers of institutional investors and managers. From an academic perspective,

the main analysis of my thesis contributes to the current stance of research by revitalizing the

deadlocked CSP and CFP debate in developed markets. First, it is the first study to provide

empirical evidence for the resource-based view of CSR in the so far untouched post-financial

crisis period of 2010-2016. Second, it is the first study that examines the link between CSR

performance and firm value across global emerging markets. It implies for future research that

CSR awareness is present across global emerging markets and that superior CSP enhances firm

value for global emerging market companies as a group. At the same time, it shows that large

regional differences in this link do exist and that there is a need for more extensive ESG data to

conduct meaningful statistical analyses on regional level.

Furthermore, it provides confirmatory evidence for the proposed link between institutions and

transaction costs proposed by El Ghoul et al. (2017) and implies that future research on CSR

should recognize that the effectiveness of CSR performance channels depends on the

institutional context of the economy where the company is located in. The additional analysis

on firm-level governance ratings implies that future research on corporate governance should

recognize that institutions, amongst other factors, effect the governance requirements of

companies and the ways in which those can be resolved. Finally, the additional analysis on the

role of sell-side analysts implies that future research should recognize that both analyst

coverage and CSR performance help reduce transaction costs and tap additional resources in

emerging markets and are complementarily valued by the market.

From a practitioner’s perspective, my findings imply that portfolio managers of institutional

investors allocating capital to emerging market equities should pick stocks of companies with

strong future CSR capabilities, large analyst coverage, and conduct firm-level governance

analyses considering PP problems rather than (relying on generic governance scores based

primarily on) PA problems and generally treat Thomson Reuters EIKON ESG scores in their

investment decisions with caution. My support for the resource-based view calls managers of

10

emerging markets companies to strive for CSR excellence to reduce transaction costs, tap

additional resources, and access desired growth capital.

This thesis is structured as follows. Section two presents the academic literature relating to the

research question at hand in more detail. Additionally, four hypotheses relating to the research

question above are developed. Section three details the sample construction process and the

methodology used to empirically test the hypotheses developed in section two. Section four

presents the descriptive statistics of the constructed sample and the inferential statistics of the

empirical tests of the four hypotheses. Furthermore, a robustness check for the baseline

hypothesis is conducted. In section 5, the statistical results are interpreted in the context of the

introduced literature and limitations of the analysis are discussed. Section 6 synthesizes the

previous sections by concluding on the implications of this study and provides motivation for

future research.

11

2. Literature Review

The literature review is structured as follows. In the first section, often-misused terminology in

the CSR realm is clarified. In the second section, an overview of the origins and subsequent

developments of the CSP and CFP debate and recent advancements in ESG data availability is

presented. In the third section, the literature on CSR in the societal, organizational, and

institutional context of emerging markets is reviewed. The fourth and last section combines the

previously presented insights on the CSP and CFP debate and the institutional context of

emerging markets to hypothesize the effect of CSR on firm value across global emerging

markets. Furthermore, it elaborates on the role of firm-level governance, sell-side analysts, and

institutional ownership on this relationship.

2.1 CSR in the context of SRI and ESG

There is an increasing trend of institutional investors acting in the long-term interests of their

beneficiaries towards investment strategies often grouped into and termed Sustainable,

Responsible and Impact (SRI) investing (USSIF, 2016). SRI investing includes the

incorporation of environmental, social, and governance (ESG) aspects in the investment

analyses and portfolio selection processes of institutional investors, next to traditional economic

factors. On top of that, institutional investors increasingly engage in active ownership, i.e. they

promote ESG amongst their portfolio companies either by informal influence or formally by

filing or co-filing of shareholder resolutions concerning ESG issues (Dimson, Karakaş, and Li,

2015). In its 2016 report, USSIF exemplarily shows the immense growth of SRI investing. US-

domiciled assets under management using SRI strategies grew by 33% from 2014 to 2016 to

USD 8.72 trillion, which is a 14-fold increase since 1995. The SRI assets in 2016 represent

nearly 22% of all tracked U.S. assets under professional management.

The increase in SRI investing can be observed not only in the U.S., but also internationally. By

signing the Principles of Responsible Investments (PRI), who were developed by institutional

investors in conjunction with the United Nations Secretary-General, investors voluntarily

commit to promote ESG aspects among all companies they are invested in. On its website it

states that the assets managed by institutional investors which committed to the PRIs amount

to around USD 80 trillion and nearly 2250 signatories spread over all continents in 2018

(UNPRI, 2018). This illustrates that ESG screening and shareholder activism towards more

12

sustainable business practices of investees has become mainstream practice. Looking at this

trend from a company’s point of view, ESG excellence has never been more vital in attracting

capital. Therefore, it is no surprise that especially large multinational enterprises (MNEs) of

both developed and developing countries have been increasingly incorporating interests of a

variety of stakeholders in their strategic decision-making processes over the last decade (Liang

and Renneboog, 2017).

Corporate social responsibility (CSR) is often defined as a mechanism to address externalities

that firms generate in the process of pursuing profit maximization that are not internalized by

shareholders (Lian and Renneboog, 2017). These externalities span over a variety of activities

contained in the environmental and social pillars of ESG, but are conceptually different from

the governance pillar. This is due to the firm-intern nature of corporate governance structures

which do not directly create externalities of society’s concern, but directly affect shareholders

(Flammer, 2013). To illustrate, firms polluting rivers nearby communities use as drinking water

(environmental) or making use of child labour in less developed countries (societal)

immediately and negatively affect the wider society. However, firms having fewer independent

directors in their board primarily concern shareholders of these firms, not society at large.

CSR definitions further pronounce that the need for companies to address externalities arises

from the economic, legal, ethical, and discretionary expectations that society places on them

(Kim et al., 2014; Liang and Renneboog, 2017). Societal expectations on firms to act

responsibly have increased significantly over time and consequently, corporate social

responsibility has increasingly become a mainstream business activity (Kitzmueller and

Shimshack, 2012). CSR activities aimed at fulfilling these expectations span over a multitude

of dimensions, such as providing employee benefits, investing in environmentally-friendly

production processes, preventing the use of child labour along the supply chain, supporting

NGOs or establishing foundations specialized in cultural and educational support in less

developed countries (e.g. Liang and Renneboog, 2017).

In short, ESG summarizes the non-financial performance dimensions of a firm about which SRI

investors care. CSR can be viewed as a subpart of ESG relating to the environmental and social

dimensions which affect society at large, while governance issues are of more firm-intern

nature. After having established an understanding of the differences in perspectives and

concepts between SRI, ESG, and CSR, I turn to the academic research on these topics. Due to

the conceptual differences in CSR and the governance categories of ESG, I separately elaborate

on these issues in the following sequences. I primarily inspect the debate that has been receiving

13

the greatest attention, which is the relationship between corporate social performance (CSP)

and corporate financial performance (CFP). In additional analyses, I further inspect the role of

internal corporate governance, the role of sell-side analysts, and the role of institutional

ownership in the relationship between CSP and CFP across global emerging markets. In the

next sequence, I provide an overview of the origins and subsequent developments of the CSP

and CFP debate.

2.2 The long-standing debate on CSP and CFP

2.2.1 CSR and stock performance

There are two general contradicting standpoints regarding the role of CSR. On the one hand,

the resource-based view states that CSR investments create necessary resources and stakeholder

support (e.g. Jones, 1995) that can translate into sustainable long-term competitive advantage

and thus higher financial performance (e.g. Waddock and Graves (1997); Russo and Fouts,

1997; Deng, Kang, and Low, 2010). On the other hand, the negative view suggests that

companies should not internalize the negative externalities they exert on other stakeholders than

shareholders, such as communities, employees, or the environment. According to them, doing

so would conflict the sole responsibility of firms, which is to maximize shareholder value (e.g.

Pigou, 1920; Friedman, 1970). Theoretical arguments through which exact channels CSR might

lead to either superior or inferior stock performance are manifold.

Proponents of the resource-based view argue that continuous stakeholder-management

relationships induced by CSR serve as monitoring and enforcement mechanism that focuses

managers on financial goals (Hill and Jones 1992; Jones 1995). This is related to the good-

governance view according to which CSR investments are a signal of properly incentivized and

governed managers who contribute to better firm performance (Ferell, Liang, and Renneboog,

2016). According to the internal resources/learning perspective CSR activities may help to build

managerial competencies because it necessitates significant employee involvement,

organization-wide coordination, and a forward-thinking managerial style (Shrivastava 1995).

Managers who acquired these capabilities are better equipped to adapt to external changes,

turbulences, and crises (Russo and Fouts, 1997). According to the reputation perspective, CSR

acts a tool to build a positive image with customers, investors, bankers, and suppliers which

facilitate their access to capital (Fombrun and Shanley 1990) and builds a reputation for quality,

reliability, and trust in the market (McWilliams and Siegel, 2001).

14

Proponents of the negative view argue that CSR expenditures are simply as a waste of scarce

resources, as they increase operating costs, represent a manifestation of agency problems, are

time-consuming and distract managers from their core responsibilities (Jensen and Meckling,

1976; Bénabou and Tirole, 2010). Empirical studies find that managers benefit at the expense

of shareholders by engaging in inappropriate CSR conduct which is self-serving as they choose

projects which earn them a good personal reputation and desired social networks among key

stakeholders, rather than supporting the firm (Krueger, 2015). Moreover, managers are inclined

to overinvest in CSR to enhance their firm’s sustainability rating to reduce the probability of

their replacements in the future (Barnea and Rubin, 2010). Other studies criticize the resource-

based causality and propose an inverse positive relationship, meaning that only well-performing

firms can afford to invest in CSR (e.g., Hong, Kubik, and Scheinkman, 2012). They reason that

deciding on CSR expenditures often represents an area of relatively high managerial discretion,

so that the initiation or cancellation of voluntary social and environmental policies depends to

a large extent on the availability of excess funds (McGuire, Sundgren, and Schneeweis, 1988).

Empirically, these early studies often compare returns of sustainable mutual funds or

sustainability indices with their conventional counterparts. However, these studies fail to

establish an academic consensus. Some studies find little support (e.g. Cummings, 2000) for

superior risk-adjusted returns, some studies contest no significant differences between them

(e.g. Sauer, 1997) and some find inferior returns for sustainable firms (e.g. Brammer, Brooks,

and Pavelin, 2006). Despite the absence of an academic consensus, the continuing above-

described trend of increasing capital commitments to SRI investment vehicles favours the view

that CSR has some form of value. Facilitated by the severe erosion of investor wealth and

numerous shutdowns of businesses during crises periods and their frequent occurrence in more

than ever interconnected global financial markets such as the recent sequence of the 1997 Asian

financial crisis, the Nasdaq internet bubble in 2000 and the global financial crisis in 2008, the

research on CSR activity shifted away from the deadlocked return debate towards the effect of

CSR on firm risk characteristics.

2.2.2 CSR and firm risk characteristics

Proponents of the resource-based view argue that – even though CSP might not directly impact

observable returns – socially responsible firms have a more favourable risk profile (Goss and

Roberts, 2011). Empirical studies find that CSR performance is on average associated with

lower idiosyncratic risk and lower probability of financial distress (Lee and Faff, 2009), lower

https://www.sciencedirect.com/science/article/pii/S0304405X14001925#bib8

15

cost of capital (e.g. El Ghoul, Guedhami, Kwok, and Mishra, 2011), lower stock price crash

risk (e.g. Kim et al., 2014; Utz, 2017), and increased analyst following and access to

institutional capital (Bushee and Noe, 2000). Two literature streams to provide rationale in

favour of this risk mitigation view of CSR evolved: higher information quality and the building

of moral capital.

Studies supporting the information quality rationale find that socially responsible firms suffer

less from earnings management, have higher financial reporting quality, less overconfident

managers, and disclose more financial information (Gelb and Strawser, 2001; Kim, Park, and

Wier, 2012; McCarthy, Oliver, and Song, 2017). Moreover, Waddock and Graves (1997) argue

that CSR investments can have a signalling effect of manager’s commitment to reduce

principal-agency conflicts which reduces the perceived risk of investors.

Studies relating to moral capital building argue that CSR investments can reduce risk exposure

through insurance-like protection in bad times by generating moral capital or goodwill among

stakeholders in good times (e.g. Godfrey, 2005; Godfrey, Merrill, and Hansen, 2009; Attig, El

Ghoul, Guedhami, and Suh, 2013). In this context, moral capital spans over a variety of

intangible internal resources such as effective employee commitment, legitimacy among

communities and regulators, trust among partners and suppliers, credibility and enhanced brand

equity among customers, and more attractiveness for investors (Godfrey, 2005). Attig, Cleary,

El Ghoul, and Guedhami (2014) support this view by finding empirically that responsible firms

are less exposed to legal, regulatory, and reputational risks and ultimately exhibit more stable

cash flows. Flammer (2013) finds that environmental commitments alleviate the consequences

of bad news event. Lins, Servaes, and Tamayo (2017) find that socially responsible firms

benefitted from higher levels of trust during the global financial crisis, translating into less

severe stock return drops, higher profitability, growth and sales per employee and better access

to debt.

Proponents of the of the negative view on CSR disagree and state that management may use

highly discretionary CSR activities to conceal firm misbehaviour, which increases financial risk

once this bad news hoarding is detected (Hemingway and Maclagan, 2004). Opposing the

information quality argument, other studies find empirical support for a positive relationship

between CSR activities and earnings management (e.g. Petrovits, 2006; Prior, Surroca, and

Tribó, 2008).

16

2.2.3 Advancements in ESG data quality/availability

Due to the increasing demand for reliable and extensive information on ESG and CSR of

investors, several data providers collect firm-level ESG information and construct scores to

make firms comparable along these non-financial performance dimensions. The ESG rating

agencies provide rating services, research, compliance and consulting services analogous to

those provided by credit rating agencies – but with a focus on ESG criteria. The three most

relevant ESG rating providers are Thomson Reuters EIKON (including formerly ASSET4),

MSCI ESG (including formerly Kinder, Lydenberg, and Domini Research & Analytics (KLD)),

and Bloomberg. Since the past decade, the ESG rating industry has been growing tremendously

and it has been consolidating. This is no surprise as addressing ESG issues has become a risk-

management concern for investors, shareholder, governments and firms and academics have

been increasingly focussing on ESG aspects (Kitzmueller and Shimshack, 2012). Thereby,

databases increase significantly in value with their coverage. Consolidating databases creates

positive synergies or might have even been necessary to survive in this data and research-driven

industry.

The emergence of more and more comprehensive ESG rating databases also changed the

academic landscape considerably. Early studies compared SRI indices, SRI investment funds,

or self-constructed ESG portfolios or funds with their conventional counterparts in terms of

risk-adjusted returns. The researchers mostly used econometrical frameworks like cross-

sectional Fama and MacBeth regressions or some Capital Asset Pricing Model (CAPM)

extension like the Carhart four-factor model. In comparison to these regression techniques with

binary ESG inclusion dummy variables, ESG rating scores have the advantage that they provide

scalable and firm-specific data, which results in large panel data sets (Halbritter and Dorfleitner,

2015). That is why most recent research published in leading journals relating to the CSR

literature stream in advanced markets relies on these ESG ratings and panel data statistics (e.g.

Kim et al., 2014; Utz, 2017). For emerging markets, the availability of large-scale data over

long periods of time is substantially less extensive as these rating providers have only been

gradually extending their coverage of firms from developed to emerging markets. However,

there are very recent papers who do exploit these advancements in ESG data availability for

individual emerging markets and construct data panels (Lee, 2016; Zhang, Xie, and Xu, 2016).

It is important to recognize that independent of the individual aspects considered and

methodologies used, there is still no consensus on the relationship between CSP and CFP.

However, there is consensus amongst academics that the extent, content, and communication

17

intensity of CSR differs significantly not only across corporations, but also across regions, and

countries (Maignan and Ralston, 2002). Therefore, before conducting an empirical study across

global emerging markets, it is essential to review the societal, organizational, and institutional

context of emerging versus developed markets.

2.3 CSR in the societal, organizational, and institutional context of emerging

markets

2.3.1 Cross-country differences in CSR activity

While CSR studies initially focussed on developed markets, emerging markets have been

increasingly receiving attention (e.g. Welford, 2004; Baughn, Bodie, and McIntosh, 2007; Qu,

2007). Still, the empirical research on the impact and relevance of corporate social

responsibility in emerging markets is still very limited (Arya and Zhang, 2009). The few early

studies are sceptical towards CSR sensitivity in emerging markets. They claim that emerging

market companies engage less in CSR activities than developed market companies due to lower

economic development levels (e.g. Welford, 2004). The reasoning usually posits that lower

income levels, less awareness of and sophistication about social and environmental problems,

lower levels of product variety, and greater emphasis on the basic value proposition of products

all contribute to less CSR sensitivity in emerging markets (e.g. Baughn et al., 2007). Moreover,

- with reference to the inverse positive relationship of CSP and CFP proposed by McGuire et.

al (1988) – they argue that emerging market firms might have less availability of excess funds

and therefore simply not the “luxury” to engage in CSR.

Research shifted subsequently towards the determinants of observed cross-country differences

in CSR activity. For example, Liang and Renneboog (2017) find that firms from common law

countries have lower CSR ratings than companies from civil law countries. They further find

that the legal origin is a stronger explanator of the cross-country variation in CSR than

previously proposed firm or country factors such as ownership concentration, political

institutions, and globalization. Attig, Boubakri, El Ghoul, and Guedhami (2016) find that for a

large sample of firms from 44 countries, firm internationalization is positively related to their

respective CSR ratings. Li, Fetscherin, Alon, Lattemann, and Yeh (2010) find that for the 105

largest MNEs in Brazil, Russia, India, and China (BRIC), a country’s governance environment

is the most important driving force behind CSR communication intensity.

18

All these studies have in common that they draw on institutional theory to explain cross-country

CSR variation. According to Hoskisson, Eden, Lau, and Wright (2000), institutional theory has

become the predominant theory for analysing management decisions in emerging markets.

Therefore, I review the literature on institutions in emerging markets next, to examine whether

there are common institutional features that characterise and distinguish global emerging

markets from global developed markets. Specifically, I provide theoretic rationale to conduct

CSR studies for global emerging markets as a group.

2.3.2 Institutional context and commonalities in emerging markets

Institutions, or specialized intermediaries in an economy emerge in response to mitigate the

various transaction costs associated with market failures in the economic exchange process

arising from negotiating, monitoring, and enforcement costs (Coase, 1937; Williamson, 1975,

1985; Jones and Hill, 1988). These institutions shape the general business environment such as

political, economic, social, legal, and technological conditions and support the effective

functioning of the market by allowing firms and individuals to trade without incurring undue

costs or risks which in turn determine the outcomes and effectiveness of organizations (Meyer

et al., 2009).

Formal institutions comprise the functioning of capital markets and the enforcement of laws,

and regulations regarding e.g. accounting requirements, information disclosure, and securities

trading. Informal institutions comprise relational ties, business groups, family connections and

government contacts (Young et al., 2008). According to leading scholars (e.g. North, 1990,

1994; Peng and Heath, 1996; Meyer et al., 2009), emerging markets across the globe – while

large cross-country differences exist – generally have less efficient formal institutions in

promoting impersonal exchanges between economic actors, resulting in business being guided

to a larger degree by informal institutions. This, in turn, has considerable consequences on the

general business environment in these markets. In emerging markets, principal-principal (PP)

conflicts between controlling shareholders and minority shareholders are more important and

pronounced rather than traditional principal-agent (PA) conflicts examined in most research

dealing with developed markets (Young et al., 2008). Agents (top managers) are also (or

represent) often the controlling shareholders via pyramid ownership structures and therewith

can circumvent monitoring mechanisms such as the board of directors (Dharwadkar, George,

and Brandes, 2000).

19

These PP conflicts manifest themselves via concentrated firm ownership through families or

the state, and often result in weak governance indicators such as fewer publicly traded firms

(La Porta, Lopez‐de‐Silanes, and Shleifer, 1999), lower firm valuations (Claessens, Djankov,

Fan, and Lang, 2002), inefficient strategy (Filatotchev, Wright, Uhlenbruck, Tihanyi, and

Hoskisson, 2003), less information contained in stock prices (Morck, Yeung, and Yu, 2000),

less investment in innovation (Morck, Wolfenzon, and Yeung, 2005). Most importantly, they

ultimately increase the risk of expropriation of minority shareholders (e.g. Claessens et al.,

2000; Faccio et al., 2001). Johnson, Boone, Breach, and Friedman (2000) find that even firms

with good reputation exploited minority shareholders during the Asian financial crisis during

the late 1990s. In this environment, it is more difficult to specify and measure the terms of

contracts as formal institutional structures are ambiguous, so transaction costs in economic

exchanges across global emerging markets are higher (Peng, 2003).

2.3.3 Strategic responses of emerging market firms to overcome institutional voids

Khanna and Palepu (1997, 2011) argue that as in emerging markets institutions like efficient

capital markets, regulatory systems, and contract enforcement mechanisms are weak, firms

must develop strategic responses to overcome these voids. In this context, Young et al. (2008)

posit that if emerging market companies seek access to minority capital they will have to incur

bonding costs as a type of implicit guarantee against expropriation of minority shareholders.

These strategic responses to bond with minority shareholders span over building a reputation

of treating minority shareholders fairly (Gomes, 2000), strategic alliances with foreign

multinationals from countries with strong institutions (Siegel, 2009), geographical clustering to

create local business environments (Karna et al., 2013), expanding the business abroad to access

more efficient and munificent foreign markets (Luo and Tung, 2007), or signalling commitment

to fair practices by voluntary cross-listings on exchanges with strong monitoring standards

(Young et al., 2008).

The very recent study of El Ghoul et al. (2017) is the first to test whether CSR might be another

type of strategic response to overcome the transaction costs associated with institutional voids.

They claim that CSR initiatives help reduce transaction cost and improve access to resources,

which creates firm value. They find supportive evidence across 53 countries during 2003-2010

that CSR is more positively related to firm value in countries with weaker market institutions.

Specifically, they find that CSR is adding value by being associated with improved access to

financing in countries with weaker equity and credit markets, greater investment and lower

20

default risk in countries with more limited business freedom, and longer trade credit period and

higher future sales growth in countries with weaker legal institutions. Linking this finding to

the fact that global emerging markets as a group generally have weaker formal market-

supporting institutions, there seems to be an increased incentive for firms in emerging markets

to engage in CSR.

2.3.4 Societal and organizational trends driving CSR sensitivity across global emerging

markets

For the channels between CSP and CFP to work, it is essential that the societies in emerging

markets care about and value CSR commitments. While initial studies argue that stakeholders

are less CSR sensitive in emerging markets or do not have the luxury to engage in CSR (e.g.

Baughn et al., 2007), I find several societal and organizational trends that mitigate the reasons

brought forward in these studies and are likely to further increase conduct and awareness of

CSR activities in emerging markets.

First, Western multinational enterprises (MNEs) increasingly focus on outsourcing production

units and expanding sales to emerging markets as a source of future growth and increased

profitability (e.g. Lacy and Hayward, 2011). Due to this increased internationalization of

supply chains, issues in health and product safety of products produced in emerging markets,

such as the milk and toy scandals in China, affect MNEs directly as their stakeholders demand

responsibility along the whole supply chain (e.g. Bogdanich, 2008). Utz (2017) for example

argues that globalization forces firms from Asia-Pacific to overinvest in CSR to adapt to

Western standards. Perez-Batres et al. (2010) find that companies of countries in Latin America

that have close business ties with European countries face more pressure on CSR issues and

thus increasingly engage in CSR activities.

Second, multinational enterprises from emerging markets (EM MNEs) increasingly emerged,

such as Embraer of Brazil, Tata of India, and Alibaba and Haier (both of China). Attig et al.

(2016) find that firm internationalization is positively related to firm’s CSR performance. As

these firms enter other emerging and developed markets, they gain considerable world-wide

attention and their activities are thus in the spotlight of investors, academics, governments, other

concerned group and individuals. Therefore, they are likely fall under the same scrutiny

regarding CSR as the traditional Western MNEs, as they face heightened awareness of these

stakeholders about pollution, product quality, and safety affecting the world at large

21

(Bogdanich, 2008) and are likely to further drive the convergence of CSR awareness in

emerging markets to Western standards (e.g. Doh, Littell and Quigley, 2015).

Third, Western SRI investors with large capital accumulations in continuously low-interest

environments amongst developed markets are searching for yield. At the same time, emerging

markets have been increasingly recognized as the main factor for international diversification

gains in portfolios of these investors (e.g. Goetzmann, Li, and Rowenhorst, 2005). According

to a study of the McKinsey Global Institute (2010), the demand for capital across emerging

markets is expected to keep surging for the foreseeable future. Emerging market companies

thus have an incentive to increase their CSR performance to pass ESG screenings to access

these capital pools via equity financing to drive their desired growth.

Fourth, strong CSR awareness of managers of emerging market firms as well as growing

education levels and middle classes facilitate CSR sensitivity in these markets. Besides, Arya

and Zhang (2009) state that social and environmental crises are usually most acutely felt in

emerging countries. Lacy and Hayward (2011) conduct a CEO study on the topic of

sustainability in partnership with United Nations Global Compact (UNGC) and Accenture. The

study is based on a survey of 766 UNGC CEOs, in-depth interviews with an additional 50

member CEOs and further interviews with more than 50 business and civil society leaders.

They find that a total of 93 percent of CEOs see sustainability as important to their company’s

future success. Furthermore, this figure is even higher in emerging markets, at 98 percent. They

recognize that it is contestable if this survey alone represents a genuine shift towards a new

approach to sustainability. However, they claim that among leading emerging market

companies, momentum seems to be building both in words and in actions.

2.4 The CSP and CFP debate in the institutional context of emerging markets

2.4.1 The effect of CSR on firm value across global emerging markets

In summary, institutional theory suggests that – even though cross-country differences exist

between them – global emerging markets as a group are typically characterised by weaker

institutions and therewith economic exchanges endure higher transaction costs. These

institutional voids shape the business environment and often induce PP conflicts via extensive

family ownership and control, business group structures, and weak legal protection of minority

shareholders and ultimately increase the risk of expropriation for minority shareholders. To

overcome these voids and reduce undue transaction costs, firms must develop strategic

22

responses that alleviate these concerns of investors. El Ghoul et al. (2017) show that superior

CSP reduces transaction costs and improves access to resources for companies in countries with

weak institutions, which is reflected in higher firm values. Furthermore, societal and

organizational trends emphasize the increase in CSR sensitivity of stakeholders and managers

of emerging market firms. Following this logic, recent research strongly points towards a

positive relationship between CSP and firm value across global emerging markets.

H1: CSR performance is positively related to firm value across global emerging markets

While the theoretical rationale to conduct CSR studies across global emerging markets is

straightforward, there is no previous study examining this link treating global emerging market

companies as a group. This is probably due to the neglection to recognize institutional

commonalities across global emerging market companies, the ESG data scarcity in emerging

markets and the extremely volatile capital markets in the last decade steering the focus on risk

characteristics. This thesis closes this research gap.

2.4.2 The role of firm-level governance in emerging markets

As institutional theory suggests, emerging market companies often lack effective market

supporting institutions, and thus suffer from a weak macro governance environment overseeing

economic transactions. Therefore, the importance investors place on efficient internal corporate

governance is especially high in emerging markets (Ding et al., 2010). Theoretically, this

suggests that firms with higher governance rating performance enjoy higher firm valuations. In

a similar study for a sample of U.S. firms, Kim et al. (2014) find that the mitigating effect of

CSR on stock price crash risk is especially pronounced when the firm-level governance rating

performance is low.

However, there is a problem arising from institutional differences between developed markets

and emerging markets. This is because the predominant model of corporate governance is a

product of developed economies, where ownership and control are often separated, and legal

mechanisms protect owners’ interests. That is why most attention when evaluating a firm’s

governance practices is placed on PA conflicts between owners and managers (Jensen and

Meckling, 1976). However, the traditional focus on PA conflicts does not apply to emerging

economies, where PP conflicts dominate the governance environment (Young et al., 2008). For

example, in developed economies concentrated ownership is widely promoted as a possible

means of addressing PA conflicts (Demsetz and Lehn, 1985). In emerging economies, however,

23

since concentrated ownership is the root cause of PP conflicts, increasing ownership

concentration even further often make things worse (Faccio et al., 2001).

Thomson Reuters – like other ESG rating agencies – has a standard criteria list used for all

companies which they evaluate. The governance pillar consists of three categories, i.e.

management, shareholders, and CSR strategy. In this context, CSR strategy refers to the

establishment of a sustainability committee, voluntarily disclosures, audits etc. Criteria for the

shareholder category are for example shareholder policy engagements and majority

requirements for director elections. While a high performance in Thomson Reuters governance

performance is in line with proper governance structures according to the developed markets

governance model, it might be less so in the institutional context of emerging markets.

Specifically, shareholders engaging in policy matters and having significant influence on the

composition of the board of directors in developed countries with dispersed ownership will

likely discipline managers. However, in emerging countries already dominant shareholders

might use their engagement to steer business according to their personal interests and might put

captured state officials or family members in the board of directors. This must not destroy value

per se, as they could potentially better manage a company in the institutional context of

emerging markets with institutional voids, but it could. Therefore, the effect of firm-level

governance rating performance on firm value remains an empirical question and calls for a

detailed inspection of all category scores.

H2: Firm-level governance rating performance has a positive effect on firm value across global

emerging markets

While firm-level corporate governance constitutes an internal monitoring mechanism in the

absence of market-supporting institutions in emerging markets, previous research has identified

two key external monitoring mechanisms which also potentially reduce the risk of expropriation

of minority shareholders, i.e. analyst coverage and institutional ownership.

2.4.3 The role of sell-side analysts on CSR and firm value across global emerging markets

According to Chung and Jo (1996), equity analysts act as an external monitor that help reduce

agency costs, and as information intermediaries who steer investor attention towards important

information. Yu (2008) claims that analyst coverage imposes discipline on misbehaving

managers and helps align managers with the interests of shareholders. The primary role of

24

equity analysts concerning governance is to uncover any financial reporting irregularities rather

than providing elaborate ESG information (Berk and DeMarzo, 2011). Consequently, firm

value should increase in the number of financial analysts following the firm – next to CSR

performance. However, in societies where economic exchanges are predominantly guided by

informal institutions such as relational ties, business groups, family connections and

government contacts (e.g. Young et al., 2008), analysing predominantly publicly available

information might not yield superior intelligence. Chan and Hameed (2006) for example find

that greater analyst coverage increases stock price synchronicity across global emerging

markets. That means that the stocks of firms covered by more analysts have less firm-specific

information content than the ones which are covered by less analysts. Therefore, whether

analyst coverage has a positive effect (next to CSR performance) on firm value remains an

empirical question.

H3: Analyst coverage has a positive effect on firm value across global emerging markets

Jo and Harjoto (2014) find for a sample of U.S. firms, that while analysts are primarily

concerned with financial information, they provide indirect but additional social pressure on

firms to reduce their irresponsible activities. Sell-side analysts work for brokerage firms and

their trading recommendations serve as basis for decisions of a large pool of clients. As such,

they work towards a reputation for accurate recommendations. In the emerging market context,

that means that they would shy away from covering companies for which there are severe

governance concerns resulting in high risk of expropriation of minority shareholder concerns –

which are their clients. As such, analyst coverage could – just like superior CSP– signal trust

to the marketplace that the covered company is well governed and bears little risk of minority

shareholder expropriation and these companies should, in turn, enjoy higher valuations.

Following this logic, both analysts as external monitors in economies with weak institutions

and CSR as internal strategic response to overcome institutional voids could be partly

substitutes for each other. In that case, analyst coverage would have a negative incremental

effect on the hypothesized positive relation between CSR and firm value.

At the same time, CSR activities – especially when intended to attract foreign capital – must be

effectively communicated to become a source of competitive advantage (Chahal and Sharma,

2006). As such, analysts could act as an amplifying voice in disseminating CSR commitments

of emerging market companies and ESG rating agencies and equity analysts would be

complementarians rather than competitors. Dimson et al. (2015) suggest that analyst coverage

25

intensifies reputational concerns along U.S. firms, which promotes the success of management

change in cases where shareholders address environmental and social concerns. In these cases,

analyst coverage would have a positive incremental effect on the hypothesized positive relation

between CSR and firm value. Which effect prevails remains an empirical question.

H4: Analyst coverage amplifies the positive effect of CSR on firm value across global emerging

markets

2.4.4 The role of institutional ownership in global emerging markets

As elaborated on before, large (Western) institutional investors increasingly exercise their rights

as business owners to influence the management of their portfolio companies to address ESG

concerns. They do so by engaging with management via letters, emails, telephone

conversations, personal meetings with senior management and voting at shareholders’ meetings

on behalf of both their internal and external clients (Dimson et al., 2015). For a sample of U.S.

public companies from 1999–2009, Dimson et al. (2015) find that successful (unsuccessful)

ESG engagements are followed by positive (zero) abnormal returns. After successful

engagements, companies experience improved accounting performance, governance and

increased institutional ownership. They further find that companies with inferior corporate

governance structures are more likely to be engaged by socially conscious institutional

investors.

Bae, Lim, and Wei (2006) argue that in companies with strong monitoring from boards or

institutional investors, adverse effects attested by proponents of the negative view of CSR such

as bad news hoarding might be limited. The investment horizon of the institutional investor is

critical. Callen and Fang (2013) find that in the U.S., the presence of long-term institutional

investors reduce crash risk by limiting managerial bad-news hoarding, but short-term

institutional investors are more likely to cause crashes through frequent trading due to high

sensitivity to bad news. Overall, the current research focussing on developed markets favours

the view that the presence of large and long-term oriented investors is associated with better

return and risk profiles of companies. There is reason to believe that the same mechanism is

less effective across emerging market companies. Dimson et al. (2015) point out that

collaboration among activists is instrumental in increasing the success rate of environmental

and social engagements. First, large institution with fiduciary duty are unlikely to have offices

in every single emerging market in which they invest and cultural differences might be large.

26

Second, emerging markets are characterised by a more informal business and governance

environment. Powerful managing principals belonging to influential families or working for the

state might be less likely to collaborate when being engaged.

As such, it would be interesting to empirically test – just like for analyst coverage – if foreign

long-term institutional shareholding is associated with higher subsequent firm valuation across

global emerging markets and if and to which extent it is a substitute or complementary

mechanism compared to CSR. Unfortunately, the data availability on institutional

shareholdings accessible on Thomson Reuters EIKON is insufficient to construct a sample of

significant size that establishes confidence in the statistical power of any empirical analysis.

Therefore, I refrain from stating explicit hypotheses regarding the role of foreign institutional

shareholdings in global emerging markets, but I encourage further researchers to address this

issue.

Figure 1 below provides an overview of the different relationships presented in this thesis and

my hypothesized interrelations of those.

Figure 1: Model of Relationships

27

3. Research Design

The research design chapter introduces the statistical set up of the empirical analysis conducted

and is structured as follows. First, a detailed overview of how the sample has been constructed

is given. Specifically, it elaborates on the sources of the raw data for dependent, independent,

and control variables with their necessary exclusions, and the construction of the proxies for

CSR performance and firm value. Second, it introduces the methodological set up of the

unbalanced panel data analysis by elaborating on econometric frameworks and the specification

used for the data at hand.

3.1 Sample construction

3.1.1 The sample

The raw data on ESG scorings, firm value, and control variables have been retrieved from

Thomson Reuters EIKON in July 2018. The creation of a global sample of companies across

emerging markets, a main distinctive aspect of my thesis, requires restructuring of the raw data.

Specifically, Thomson Reuters EIKON provides ESG, market and fundamental data per

financial year. However, closer examination reveals that these financial years differ between

companies in terms of reporting date within a year and in terms of data availability. This means

for example that the total value of common book equity of company x for the past 5 years spans

over the time horizon of 2012-2016 and it reports on June 30 of each year, while the total value

of common book equity of company y for the past 5 years spans over the time horizon of 2013-

2017 and it reports on December 31 of each year. Consequently, the raw data had to be

restructured, so that the financial years are consistent across all companies.

Table 1 below shows an overview over the sample selection process. The starting point is the

Thomson Reuters Global Emerging Market index, which consists of 3091 companies, whose

headquarters are based in countries defined as emerging markets by Thomson Reuters. Of these

3091 companies, only around a third (950) has received an ESG rating at least once in the period

of 2010-2016. Furthermore, I exclude 13 company duplicates. There is no determining rule but

considerable controversy regarding how to classify countries as emerging. In the academic

context, an often-used benchmark when talking about emerging markets is the MSCI EM Index.

Therefore, I decide to follow this index classification scheme (Appendix A) and exclude another

28

53 companies whose headquarters are based in countries which this benchmark index does not

classify as emerging markets.

Table 1: Construction of the unbalanced data panel

In panel a of this table, the sample selection process is described. The raw data sourced from Thomson Reuters

EIKON comprise 3091 companies comprised in the Thomson Reuters Global Emerging Market Index. 657

companies, for which at least in one financial year between 2010 and 2016 full data is available, are identified and

included in the sample. The criteria in the first column involve Return on Assets (ROA), which is calculated as Net income before extraordinary items𝑖,𝑌

Total Book Value of Assets𝑖,𝑌−1, SIZE, calculated as 𝑙𝑛 (𝐵𝑜𝑜𝑘 𝑉𝑎𝑙𝑢𝑒 𝑜𝑓 𝐴𝑠𝑠𝑒𝑡𝑠𝑖,𝑌) and Tobin’s q (TOBQ),

which is calculated as 𝑀𝑎𝑟𝑘𝑒𝑡 𝐶𝑎𝑝𝑖𝑡𝑎𝑙𝑖𝑧𝑎𝑡𝑖𝑜𝑛𝑖,𝑌+𝐵𝑜𝑜𝑘 𝑉𝑎𝑙𝑢𝑒 𝑜𝑓 𝐴𝑠𝑠𝑒𝑡𝑠𝑖,𝑌−𝐵𝑜𝑜𝑘 𝑉𝑎𝑙𝑢𝑒 𝑜𝑓 𝐸𝑞𝑢𝑖𝑡𝑦𝑖,𝑌

𝐵𝑜𝑜𝑘 𝑉𝑎𝑙𝑢𝑒 𝑜𝑓 𝐴𝑠𝑠𝑒𝑡𝑠𝑖,𝑌. Panel b illustrates the

number of firm-year observations available per financial year. As this number varies over the sample period 2010-

2016, this data panel is called unbalanced.

Panel A: Overview of the unbalanced panel data sample construction

Data source/operations conducted # Companies left # Exclusions

TR Global Emerging Market Index 3091

TR Global Emerging Markets index with at least one ESG

scoring

950 2141

Excluding duplicates 937 13

Excluding non-MSCI EM defined countries 884 53

Excluding negative Book value of Equity 867 17

Trim SIZE and ROA at top and bottom percent 835 32

Excluding TOBQ >8 outliers 825 10

Excluding financial sector companies 657 168

Panel B: Number of final firm-year observations per financial year

Financial Year # Companies included in the sample

2010 424

2011 470

2012 517

2013 551

2014 595

2015 618

2016 625

TOTAL 3800

Furthermore, I exclude 17 firms with negative book values of equity as they are subject to

different bankruptcy laws and might distort the results. Histograms reveal that there several

extreme values in size and return on assets (ROA) in the 867 companies left, so I further trim

the data along these two dimensions at the bottom and top percentile. They further reveal that

there are several extreme positive outliers in Tobin’s q (TOBQ), so I delete another 10

companies which have a value higher than 8. Finally, in line with previous research (e.g. Kim

et al., 2014), I exclude 168 of the remaining companies which are classified as either banks or

insurances according to the global industry classification standards (GICS) developed by MSCI

29

and Standard & Poor's. Financial institutions are often subject to different reporting standards

and might distort the results when included. The final sample consists of 657 firms. The

increasing data availability of emerging market companies becomes evident in panel b of Table

1, which shows that the number of firm year observations per financial year is strictly increasing

in time. The number of firm-year observations across all years amounts to 3800.

Table 2 below shows a more detailed overview over the sampling distribution by country, MSCI

EM region and GICS industry group. The distribution by country ranges from merely one

company in the Czech Republic to 100 companies in Taiwan. Generally, data availability seems

to be most extensive in Asian countries and least extensive in countries of the Middle East and

Eastern Europe. This can be further seen in the sample distribution by MSCI region. Companies

clustered into the emerging market region Asia represent nearly 60% of the sample, while the

other two regions EMEA and Americas only represent around 20% each. However, there is no

reason to be concerned about systematic sampling bias. The latest MSCI EM index fact sheet

of August 2018 reveals that China, Taiwan, and India make up over 50% of the index, while

these three countries together amount to less than 40% in my sample. Instead of sample bias,

there is simply a dominance of Asian economies that are emerging. Furthermore, there are over

one hundred companies in each region, so the sample is well diversified. Panel c shows the

distribution by the remaining 22 GICS industry groups, after banks and insurances have been

excluded. While commercial & professional services and diversified financial (0.30% and

0.61%, respectively) are underrepresented, materials and capital goods (15.22% and 10.35%,

respectively) are slightly overrepresented.

This is no surprise, given that the two usually large tertiary industry sectors banks and

insurances have been excluded and the primary and secondary economic sectors in emerging

economies are typically larger compared to developed economies. Moreover, the shares of the

remaining industry groups are reasonably close to each other, so that this dimension of the

sample does also not raise serious concerns about any systematic sample bias.

30

Table 2: Sample distribution by country, region, and industry group

This table shows the sample distribution of 657 individual companies retrieved from Thomson Reuters EIKON

that are in countries defined as emerging markets according to the Morgan Stanley Capital International (MSCI)

Emerging Market (EM) Index classification during 2010-2016. Panel a shows the number and percentage of the

657 companies belonging to each of the 20 different emerging markets of this sample. Panel b shows the number

and percentage of the 657 companies belonging to each of the three regions in which the MSCI EM index is

divided. Panel c shows the number and percentage of the 657 companies belonging to each of the remaining 22

global industry classification standards (GICS) industry groups developed by MSCI and Standard & Poor's, after

the two industry groups industries banks and insurances have been excluded due to regulatory reporting

differences.

Panel A: By Country Panel C: By GICS Industry Group

Country N % GICS Industry Group N %

South Africa 75 11.42% Materials 100 15.22%

Malaysia 38 5.78% Food, Beverage & Tobacco 46 7.00%

Hong Kong 30 4.57% Capital Goods 68 10.35%

Taiwan 100 15.22% Food & Staples Retailing 17 2.59%

Thailand 30 4.57% Real Estate 49 7.46%

Philippines 22 3.35% Energy 47 7.15%

Indonesia 29 4.41% Utilities 61 9.28%

India 71 10.81% Technology Hardware & Equipment 38 5.78%

Czech Republic 1 0.15% Automobiles & Components 24 3.65%

Russia 25 3.81% Pharma, Biotechnology & Life Sciences 17 2.59%

Turkey 18 2.74% Retailing 19 2.89%

Hungary 3 0.46% Transportation 46 7.00%

Brazil 56 8.52% Semiconductors & Equipment 13 1.98%

China 62 13.44% Telecommunication Services 41 6.24%

Mexico 31 4.72% Consumer Services 14 2.13%

Egypt 5 0.76% Commercial & Prof. Services 2 0.30%

Chile 31 4.72% Household & Personal Products 5 0.76%

UA Emirates 6 0.91% Consumer Durables & Apparel 18 2.74%

Qatar 7 1.07% Diversified Financials 4 0.61%

Poland 17 2.59% Media 10 1.52%

TOTAL 657 100% Health Care Equipment & Services 10 1.52%

Panel B: By MSCI EM Region Software & Services 8 1.22%

MSCI EM Region N % TOTAL 657 100%

Asia 382 58.14%

EMEA 157 23.90%

Americas 118 17.96%

TOTAL 657 100%

3.1.2 Measuring CSR performance

For international CSR studies and especially studies examining CSR in emerging markets, the

preferred database for ESG data is Thomson Reuters EIKON, as it is one of the most

comprehensive ESG databases in the industry covering over 7,000 public companies globally,

across more than 400 different ESG metrics, increasingly enhancing the coverage of emerging

market companies since 2010 (Halbritter and Dorfleitner, 2015; Thomson Reuters, 2018).

Following these studies and given that this thesis is conducted across global emerging markets,

I rely on these Thomson Reuters EIKON ESG scores in the statistical analysis. These scores

are updated financial yearly and even include firms after bankruptcy, a merger, and other causes

31

of delisting. Thus, the data set is free from survivorship bias (Utz, 2017). They span over ten

category scores, which are combined into three pillars: environmental, social, and governance.

These are further aggregated into an ESG score and any controversies across these categories

are captured in a separate ESG controversies score. The ESG score is eventually combined with

the ESG controversies score to yield the ESG combined score. All scores range from 0-100 and

are assessed relative to their industry peers. A detailed overview of the methodology of the

Thomson Reuters EIKON ESG ratings can be found in Appendix B.

In line with the rationale to conceptually differentiate between ESG and CSR above, authors of

recent studies on CSR in international or emerging market settings across top rated journals

refer to the common practice of using the equally-weighted average of the environmental and

social pillar scores as proxy for CSR performance (e.g. Kim et al., 2014; El Ghoul et al., 2017,

Utz, 2017). Following their logic, I define CSR performance 𝐶𝑆𝑅𝑖,𝑌 of firm i in financial year

Y as:

(1) 𝐶𝑆𝑅𝑖,𝑌 = 𝐸𝑛𝑣𝑖𝑟𝑜𝑛𝑚𝑒𝑛𝑡𝑎𝑙 𝑝𝑖𝑙𝑙𝑎𝑟 𝑠𝑐𝑜𝑟𝑒𝑖,𝑌 + 𝑆𝑜𝑐𝑖𝑎𝑙 𝑝𝑖𝑙𝑙𝑎𝑟 𝑠𝑐𝑜𝑟𝑒𝑖,𝑌

2

However, this approach is subject to debate. The governance scores are composed of the

management, shareholders, and CSR strategy categories. However, CSR activity is exactly

what researchers are intending to proxy by the combination of the environmental and social

pillars. In line with this criticism, Utz (2017) and Attig et al. (2013) observe that the significance

of pillar scores in their analyses is driven by just a few categories and that some categories in

the same pillars have contradicting effects and might therefore mask statistically consistent

results at pillar level. Therefore, I conduct an additional analysis where I inspect the effect of

the individual ESG category scores on firm value.

3.1.3 Measuring firm value

A common proxy used for firm value is Tobin’s q ratio (e.g. Gompers et al., 2003; Bebchuk

and Cohen, 2005; Cremers and Nair, 2005). As CSR can translate into superior firm

performance through various channels effecting return and risk dimensions, I use Tobin’s q as

firm value proxy for financial performance, as it captures all these channels in aggregate (El

Ghoul et al., 2017). Tobin’s q (TOBQ) is defined as the market value of the firm’s assets divided

by the replacement value of the firm’s assets and indicates by how much more (less) the market

32

values the company’s assets above (below) the mere book value of the assets the firm owns. In

line with El Ghoul et al. (2017), I compute 𝑇𝑂𝐵𝑄𝑖,𝑌 for firm i in financial year Y as follows:

(2) 𝑇𝑂𝐵𝑄𝑖,𝑌 = 𝑀𝑎𝑟𝑘𝑒𝑡 𝐶𝑎𝑝𝑖𝑡𝑎𝑙𝑖𝑧𝑎𝑡𝑖𝑜𝑛𝑖,𝑌 + 𝐵𝑜𝑜𝑘 𝑉𝑎𝑙𝑢𝑒 𝑜𝑓 𝐴𝑠𝑠𝑒𝑡𝑠𝑖,𝑌 − 𝐵𝑜𝑜𝑘 𝑉𝑎𝑙𝑢𝑒 𝑜𝑓 𝐸𝑞𝑢𝑖𝑡𝑦𝑖,𝑌

𝐵𝑜𝑜𝑘 𝑉𝑎𝑙𝑢𝑒 𝑜𝑓 𝐴𝑠𝑠𝑒𝑡𝑠𝑖,𝑌

Using TOBQ as dependent variable rather than stock returns has the advantage that within the

Thomson Reuters EIKON database, market capitalization and accounting data are both

measured on the reporting date of each company and thus automatically match. Proxies

involving stock return data would necessitate the matching of trading days to the individual

financial reporting dates of each company and to guarantee comparability, trading days between

companies that are listed on different exchanges would have to be matched. Moreover,

Thomson Reuters does not report the dates at which they publish their ESG ratings. This results

in a timing problem regarding stock return and CSR performance.

3.1.4 Control variables

I follow the previously introduced studies of Gompers et al. (2013) and El Ghoul et al. (2017),

which have identified a variety of control variables expected to significantly influence TOBQ

beyond CSR. (1) Firms that are more profitable have more scope to pay dividends, invest in

necessary R&D or other projects and to sustain adverse market events and are thus expected to

have higher firm values. I define 𝑅𝑂𝐴𝑖,𝑌 as net income before extraordinary items during

financial year Y scaled by previous financial year’s total book value of assets. (2) Larger firms

are already established and are associated with lower TOBQ because these firms tend to have

more limited growth opportunities. I define 𝑆𝐼𝑍𝐸𝑖,𝑌 as the logarithm of total book value of

assets at the end of the financial year. (3) Firms with higher leverage are expected to be more

sensitive to market shocks and to have less flexibility in making investment decisions due to

for example constraining debt covenants, resulting in lower TOBQ. I define 𝐿𝐸𝑉𝑖,𝑌 as total book

value of debt at the end of the financial year over total book value of assets at the end of the

same financial year.

(4) In developed markets and international studies, economic development has been found to

be positively related to TOBQ (e.g. El Ghoul et al., 2017). The rationale is that economic

development is associated with better institutions like equity and debt markets and legal

protection, which results in higher valuations. On the other hand, countries with higher

33

economic development often yield less growth opportunities as these markets are typically more

saturated. In a purely emerging market companies sample like in this study, the effect remains

to be seen. I define 𝐺𝐷𝑃𝑖,𝑌 as the logarithm of GDP per capita each year at constant 2000 USD,

which I retrieved from the publicly available World Bank database. (5) In the institutional

environment of emerging markets, firms that have had more time to establish trust among

minority shareholders are likely to be able to attract more local and foreign capital. Therefore,

firm age is expected to be positively related to TOBQ. I follow Gompers et al. (2003) in defining

𝐴𝐺𝐸𝑖,𝑌 as the logarithm of firm age at the end of the financial year. An overview of all variables

with precise Thomson Reuters definitions can be found in Appendix C.

3.2 Methodology

3.2.1 Panel data analysis

The key feature of a data panel set is that it contains observations of individuals across both,

the cross-sectional and the time-series (longitudinal) dimensions. Therefore, panel data

regressions differ from a regular time-series or cross-section regressions in that the variables

have double subscripts. In its simplest form, the baseline regression inspected looks as follows:

(3) 𝑇𝑂𝐵𝑄𝑖,𝑌+1 = 𝛼0 + 𝛽1 ∗ 𝐶𝑆𝑅𝑖,𝑌 + 𝛽2 ∗ 𝑅𝑂𝐴𝑖,𝑌+𝛽3 ∗ 𝐿𝐸𝑉𝑖,𝑌 + 𝛽4 ∗ 𝑆𝐼𝑍𝐸𝑖,𝑌

+ 𝛽4 ∗ 𝐺𝐷𝑃𝑖,𝑌 + 𝛽4 ∗ 𝐴𝐺𝐸𝑖,𝑌 + 𝜀𝑖,𝑌

Where i denotes the individual firms (cross-section component) and Y denotes the different

financial years (time-series component) for which these individual firm observations are

collected. This baseline regression tests the effect of CSR performance on one-year ahead firm

valuation proxied by TOBQ while controlling for fundamental variables explained above. In

line with El Ghoul et al. (2017), I use lagged CSR and fundamental data to mitigate concerns

about reverse causality and simultaneity. The cross-sectional observations are the 657

individual companies. The time-series observations are regularly spaced over a period of seven

individual financial years, i.e. 2010-2016 for the independent variables and 2011-2017 for the

dependent variable.

Data panels can be either balanced or unbalanced. A balanced panel sample in this context

would mean that I only include companies in the sample that have full availability of data across

the entire time horizon of 2010-2016 for independent variables and equivalently 2011-2017 for

the dependent variable. An unbalanced panel data includes firms for which all variables are

34

available in one year, independent of whether they are available in any of the other years. The

main advantage of a balanced data panel is the statistical ease of analysing it with standard

statistical software. The disadvantage is that it necessitates the exclusion of a relatively large

amount of observations. Compared to similar studies, the sample size of 3080 firm-year

observations is on the lower end, so excluding even more observations is not desirable. More

importantly, only analysing companies with a complete history of data may inflict attrition bias

on the analysis, which relates to systematic non-response or dropping of observations which

inflict the statistical validity of the analysis (Baltagi, 2005). Specifically, there might be an

underlying reason why companies have full histories of ESG data, and only including those

companies might influence the outcome of the empirical analysis leading to falsely generalising

results only applying to this specific type of companies. For these reasons, I rely on an

unbalanced panel data regression analysis. The next section elaborates on the exact model

specification used.

3.2.2 Fixed effects OLS model

There are three different kinds of OLS panel data regression models which one needs to estimate

to subsequently select the most appropriate one. The pooled regression model pools all 3010

firm-year observations together, combining all 657 companies and therewith neglecting the

cross section and time series nature of the data. The fixed effects model allows for heterogeneity

or individuality among the cross-sections, periods or both by allowing them to have individual

time-invariant and/or period individual-invariant intercepts and error terms. The random effects

model maintains the cross-section and time series nature of the data model, but instead of

systematic individuality among these dimensions, it imposes a common mean value for the

intercept across all cross-sections and/or periods (Baltagi, 2005).

I run a pooled panel regression as a basis to conduct the Breusch and Pagan (1979) test for

heteroscedasticity and random coefficient variation. Its null hypothesis is that there is no

heteroskedasticity along both dimensions, cross-sections and periods. For both dimensions, this

is clearly rejected, providing evidence of heteroskedasticity along both dimensions at the 1%

level (Appendix D). This first attempt suggests a two-way error component unbalanced OLS

regression model as proposed by El Ghoul et al. (2017) with fixed effects for cross-sections and

periods, which looks as follows:

35

(4) 𝑇𝑂𝐵𝑄𝑖,𝑌+1 = 𝛼0 + 𝛽1 ∗ 𝐶𝑆𝑅𝑖,𝑌 + 𝛽2 ∗ 𝑅𝑂𝐴𝑖,𝑌+𝛽3 ∗ 𝐿𝐸𝑉𝑖,𝑌 + 𝛽4 ∗ 𝑆𝐼𝑍𝐸𝑖,𝑌

+ 𝛽4 ∗ 𝐺𝐷𝑃𝑖,𝑌 + 𝛽4 ∗ 𝐴𝐺𝐸𝑖,𝑌 + 𝐹𝐼𝑋𝐸𝐷 𝐸𝐹𝐹𝐸𝐶𝑇𝑆 + 𝑢𝑖,𝑌

Where 𝐹𝐼𝑋𝐸𝐷 𝐸𝐹𝐹𝐸𝐶𝑇𝑆 represent dummy variables for companies and years and the model

contains two-way error component disturbances:

𝑢𝑖,𝑌 = 𝜇𝑖,𝑌 + 𝜆𝑖,𝑌 + 𝜗𝑖,𝑌

The model error term 𝑢𝑖,𝑌 is composed of 𝜇𝑖,𝑌, which denotes the time-invariant unobservable

cross-sectional individual effect, λt denotes the individual-invariant unobservable time-series

effect and 𝜗𝑖,𝑌 is the remainder stochastic disturbance term. The critical assumption for this

two-way fixed effects error component model with ordinary least squares (OLS) estimators to

produce consistent estimators is that the remainder stochastic error term is normally distributed

with constant variance, i.e. 𝜗𝑖,𝑌 ∼IID(0,𝜎𝜗2). If the unobserved heterogeneity contained in the

remainder stochastic error is correlated with one or more of the explanatory variables, OLS

parameter estimates are biased and inconsistent (Baltagi, 2005).

3.2.3 Generalized least square regression (EGLS)

I conduct a White (1980) test for heteroskedasticity, i.e. I regress the remainder residuals of the

fixed effects model against the squared regressors. The joint significance of the explanatories

is given by the reported F-statistic of 84.621 with a corresponding p-value of 0.000 (Appendix

E). Consequently, the hypothesis that the explanatories are insignificant is rejected, so the OLS

fixed effects model suffers from heterogeneity in the remainder error term. Furthermore, a

Jarque-Bera test for normality of the standardized remainder disturbances yields a test statistic

of 7509 and is rejected at the 1% level (Appendix E). Consequently, OLS estimates are

inconsistent for our data and it is necessary to switch to a generalized least squares (GLS) model

(Baltagi, 2005).

I repeat the White test separately for cross-sectional fixed effects and period fixed effects and

find that the heteroskedasticity stems almost exclusively from the cross-section error terms.

Therefore, I specify an estimated generalized least square regression (EGLS) with cross-section

weights, as it relaxes the assumption of homoskedasticity in the cross-sections but copes with

the heteroskedasticity through a cross-sectionally weighted error term, as first suggested by

Mazodier and Trognon (1978). With this model specification, the heteroskedasticity problem

is eliminated, as the subsequent White test in Appendix F shows. Furthermore, this EGLS

36

model reveals that the standard errors of the estimators are smaller, which provides additional

evidence that for the dataset GLS is more efficient than OLS. For the research setting at hand,

this pooled GLS model specification also has a conceptual advantage over the fixed effects

methodology applied by El Ghoul et al. (2017). While they specify numerous institutional

variables to trace out specific institutional country-level determinants that effect the relationship

between CSR and firm valuation, the crucial point of my analysis is the assumption that global

emerging markets as a pooled group are determined by weak institutions.

By selecting cross-section weights in the EViews mask, the software automatically estimates a

feasible GLS specification correcting for the presence of cross-section heteroskedasticity.

Moreover, I report corresponding robust standard errors adjusted for the presence of

heteroskedasticity in the cross-sections. The final model looks as follows:

(5) 𝑇𝑂𝐵𝑄𝑖,𝑌+1 = 𝛼0 + 𝛽1 ∗ 𝐶𝑆𝑅𝑖,𝑌 + 𝛽2 ∗ 𝑅𝑂𝐴𝑖,𝑌 + 𝛽3 ∗ 𝐿𝐸𝑉𝑖,𝑌 + 𝛽4 ∗ 𝑆𝐼𝑍𝐸𝑖,𝑌

+ 𝛽4 ∗ 𝐺𝐷𝑃𝑖,𝑌 + 𝛽4 ∗ 𝐴𝐺𝐸𝑖,𝑌 + 𝛾𝑖,𝑌

with cross-sectionally weighted error disturbances:

𝛾𝑖,𝑌 = 𝑤𝑖,𝑌 ∗ 𝜇𝑖,𝑌 + 𝜗𝑖,𝑌

where 𝑤𝑖,𝑌 represents the weight placed on each cross-section residual when minimizing the

overall sum of squared residuals of the overall GLS model.

To test for regional differences in this relationship, I iteratively run this baseline regression

including the respective interaction terms CSR*Asia, CSR*EMEA, and CSR*Americas. To

test the effect of firm-level governance and ESG ratings on pillar and category level on firm

value across emerging markets, I adjust the independent variables of baseline regression (5)

accordingly. Specifically, I substitute the CSR proxy for (1) the ESG combined score, (2) the

three pillar scores (3) the governmental pillar score separately, (4) the governmental pillar

category scores: Management, Shareholders, and CSR Strategy, (5) and all 10 category scores:

Resource Use, Emissions, Environmental Innovation, Workforce, Human Rights, Community,

Product Responsibility, Management, Shareholders, CSR Strategy. The methodology,

however, does not change. The same holds true for the analysis of the effect of analyst coverage

on the relationship of CSR and firm value. Specifically, I add 𝐴𝑁𝐴𝑖,𝑌 as defined before to

baseline regression model (5). Furthermore, I add the interaction term 𝐶𝑆𝑅𝑖,𝑌*𝐴𝑁𝐴𝑖,𝑌 to inspect

whether and to which extent ESG rating providers and analysts are complementarians or

substitutes. As the number of analysts is only available for slightly less companies, the analysis

on the role of analysts is conducted on a reduced sample with 3,504 firm-year observations.

37

4. Results

This chapter provides an objective analysis of the descriptive and inferential statistics produced

by the empirical analysis and is structured as follows. First, descriptive statistics for the full

sample and by country and region are presented. Second, the results of the inferential statistics

regarding the main analysis of the effect of CSR on firm value and the additional analyses

regarding the role of firm-level governance and sell-side analysts are presented. The description

of the statistical results provides the basis for a subjective interpretation of them in the context

of the introduced literature, which follows in the fifth chapter.

4.1 Descriptive Statistics

Table 3 below shows the descriptive statistics of the unbalanced data panel. Even though I

excluded 10 outliers with values for TOBQ above 8, the dispersion in firm value is considerable.

In the international sample of 2445 firms from 53 countries over the period 2003-2010 across

both developed and emerging markets of El Ghoul et al. (2017), the mean value of TOBQ is

1.76 with a standard deviation of 0.99. In my sample, consisting of companies exclusively from

emerging markets, the mean value of TOBQ is slightly lower with 1.661 and the standard

deviation slightly higher with 1.080. Their sample yields an average value of CSR of 52.72 with

a standard deviation of 28.89. In my sample, the average value of CSR is slightly lower with

46.794 and the standard is notably lower with 20.898. Even though the sample periods differ,

and the samples span over different geographies, the values are relatively similar. The signs and

magnitudes of the control variables are also in line with previous research.

Firm size is measured as the logarithm of total book value of assets and transformation to

naturals numbers shows that it varies from around USD 140mn to around USD 410bn with a

mean and median value of around USD 5.3bn. This indicates that there are rather large

companies included in the final sample, which is probably caused by the fact that Thomson

Reuter covers rather large companies about which there is enough public information available

to arrive at accurate rating results. Moreover, the demand for ESG data of clients is likely to be

larger for large, publicly traded companies. As such, this constitutes a potential limitation in the

way that the obtained results and conclusions might exclusively apply to relatively large firms.

38

Table 3: Descriptive statistics for the full and reduced sample

This table presents the descriptive statistics for the full sample of 657 individual emerging market firms with 3,800

firm-year observations included in the baseline unbalanced data panel as well as for the reduced sample of 3,504

firm-year observations used for the additional analysis of the role of analysts (last row). One-year ahead Tobin’s

q (TOBQ) is measured for the financial reporting years from 2011-2017 and is defined as 𝑀𝑎𝑟𝑘𝑒𝑡 𝐶𝑎𝑝𝑖𝑡𝑎𝑙𝑖𝑧𝑎𝑡𝑖𝑜𝑛𝑖,𝑌+𝐵𝑜𝑜𝑘 𝑉𝑎𝑙𝑢𝑒 𝑜𝑓 𝐴𝑠𝑠𝑒𝑡𝑠𝑖,𝑌−𝐵𝑜𝑜𝑘 𝑉𝑎𝑙𝑢𝑒 𝑜𝑓 𝐸𝑞𝑢𝑖𝑡𝑦𝑖,𝑌

𝐵𝑜𝑜𝑘 𝑉𝑎𝑙𝑢𝑒 𝑜𝑓 𝐴𝑠𝑠𝑒𝑡𝑠𝑖,𝑌. All other variables are constructed with financial

yearly data ranging from 2010-2016. The Corporate Social Responsibility (CSR) proxy is constructed from

Thomson Reuters EIKON ESG scores and defined as 𝐸𝑛𝑣𝑖𝑟𝑜𝑛𝑚𝑒𝑛𝑡𝑎𝑙 𝑝𝑖𝑙𝑙𝑎𝑟 𝑠𝑐𝑜𝑟𝑒𝑖,𝑌+𝑆𝑜𝑐𝑖𝑎𝑙 𝑝𝑖𝑙𝑙𝑎𝑟 𝑠𝑐𝑜𝑟𝑒𝑖,𝑌

2. Return on

Assets (ROA) is defined as Net income before extraordinary items𝑖,𝑌

Total Book Value of Assets𝑖,𝑌−1. Leverage (LEV) is defined as

Total Reported Value of Debt𝑖,𝑌

Book Value of Assets𝑖,𝑌. SIZE is defined as the logarithm of 𝐵𝑜𝑜𝑘 𝑉𝑎𝑙𝑢𝑒 𝑜𝑓 𝐴𝑠𝑠𝑒𝑡𝑠𝑖,𝑌. GDP is defined as the

logarithm of GDP per capita in constant 2010 USD𝑖,𝑌. Age is defined as the logarithm of current Year 𝑌𝑖 −𝑌𝑒𝑎𝑟 𝑜𝑓 𝑖𝑛𝑐𝑜𝑟𝑝𝑜𝑟𝑎𝑡𝑖𝑜𝑛𝑖 . For the additional analysis of firm-level governance, the Thomson Reuters ESG

combined score, the three pillar level scores: Environmental, Social, and Governance as well as the ten category

level scores: Resource Use, Emissions, Environmental Innovation, Workforce, Human Rights, Community,

Product Responsibility, Management, Shareholders, and CSR Strategy, are included. All scores range from 0-100

and are assessed relative to their industry peers. A detailed overview of the methodology of the Thomson Reuters

EIKON ESG ratings can be found in Appendix B. For the additional analysis of analyst coverage, ANA is defined

as the logarithm of the 𝑊𝑒𝑖𝑔ℎ𝑡𝑒𝑑 − 𝑎𝑣𝑒𝑟𝑎𝑔𝑒 𝑛𝑢𝑚𝑏𝑒𝑟 𝑜𝑓 𝑠𝑒𝑙𝑙 − 𝑠𝑖𝑑𝑒 𝑎𝑛𝑎𝑙𝑦𝑠𝑡𝑠𝑖,𝑌 that are covering the stock of

the company. The raw data have been retrieved for companies in the Thomson Reuters EIKON Global Emerging

Market Index which have their headquarters based in a country defined as emerging market by the Morgan Stanley

Capital International (MSCI) Emerging Market (EM) Index classification. Next to firms with insufficient data

availability, firms with negative Book value of Equity, TOBQ >8, and financial sector firms (GICS: Banks and

Insurances) have been excluded. The sample is further trimmed at the top and bottom percent across SIZE and

ROA.

Variable N Mean Min Q1 Median Q3 Max SD

Dependent Variable

TOBQ 3800 1.661 0.323 1.019 1.273 1.887 7.889 1.080

Independent Variable

CSR (raw) 3800 46.794 5.241 28.696 47.033 63.471 94.928 20.898

CRS (normalised) 3800 0.463 0 0.262 0.466 0.649 1 0.233

Control Variables

ROA 3800 0.068 -0.317 0.027 0.054 0.096 0.880 0.073

LEV 3800 0.272 0.000 0.149 0.271 0.374 0.811 0.160

SIZE 3800 22.394 18.755 21.548 22.396 23.229 26.736 1.226

GDP 3800 9.081 7.205 8.652 9.208 9.609 11.194 0.861

AGE 3800 3.537 0.693 3.091 3.497 3.970 5.176 0.571

Additional analysis: Firm-level governance

ESG combined score 3800 44.271 6.133 31.107 42.702 56.941 92.313 17.043

Environmental Pillar (E) 3800 46.221 2.600 27.568 45.562 63.706 98.462 22.286

Social Pillar (S) 3800 47.368 3.413 27.439 47.922 65.714 96.674 22.776

Governance Pillar (G) 3800 49.567 2.605 32.998 49.200 65.942 97.506 21.054

G (normalized) 3800 0.495 0 0.320 0.491 0.667 1 0.222

Resource Use 3800 46.576 0.167 23.333 45.565 67.724 99.838 26.473

Emissions 3800 45.931 0.168 21.077 45.326 69.048 99.775 28.378

Environmental Innovation 3800 46.180 0.177 25.294 41.100 67.373 99.819 26.543

Workforce 3800 51.426 0.162 26.744 52.705 76.173 99.775 28.633

Human Rights 3800 47.723 7.692 26.800 34.574 71.104 99.746 25.980

Community 3800 39.578 0.162 14.085 32.438 64.073 99.825 29.202

Product Responsibility 3800 46.765 0.234 20.940 44.186 72.099 99.838 29.106

Management 3800 50.681 0.355 25.944 51.087 75.754 99.624 28.644

Shareholders 3800 48.547 0.021 24.428 47.442 72.642 99.645 28.390

CSR Strategy 3800 52.140 0.391 27.166 52.000 76.667 99.624 28.162

Additional analysis: Sell-side analysts

ANA 3504 2.431 0 2.079 2.565 2.944 4.025 0.788

39

The same transformation reveals that firm-age varies from just 2 years of existence to 177 years

of existence with a mean age of 35 years. This further shows that firms are not just relatively

large but also relatively mature, even though there are some very recently established firms in

the left tail of the distribution.

The ESG rating variables for the additional analysis regarding the role of firm-level governance

vary from mean values of 39.578 in the community dimension to 52.140 in the CSR strategy

dimension. Of all variables, the controversial CSR strategy category of the governance pillar,

which the common CSR proxy does not capture, scores the highest in this sample. This

emphasizes the relevance of a further investigation of the reasonableness of this commonly used

proxy. Furthermore, it is striking that the dispersion of the scores decreases in the consolidation

of categories to pillars and ultimately the ESG-combined score. This could indicate that some

important information contained in the category scores are smoothed out when combined into

pillar scores and investors merely relying on pillar or combined scores might fail to notice this

information. In line with El Ghoul et al. (2017), after pointing out these features of the raw ESG

scores, I normalise all scores for the further analysis, so that they range between 0 and 1.

However, table 3 only illustrates them for the key independent variables, CSR and Governance

Pillar, to preserve the readability of this thesis.

Data regarding the number of sell-side analysts covering the company stock is only available

for 638 individual firms with 3504 firm-year observations. ANA is also measured on a

logarithmic scale. Transforming these values shows that the weighted-average number of sell-

side analysts covering the stock per financial year varies greatly between 1 and 56, with the

average company being relatively extensively followed by 15 analysts. The relatively large

average of analyst coverage is no surprise given the previous findings that the firms included

in the final sample are relatively large and mature.

40

Table 4 below shows the mean values and the corresponding standard deviations for the two

key variables of the analysis, TOBQ and CSR performance, for each country in panel a and

each region in panel b. Panel a illustrates that the mean values for TOBQ range from 0.980 with

a standard deviation of 0.210 in Egypt to 2.445 with a standard deviation of 1.824 in India.

Interestingly, companies from Egypt also have the weakest mean value in the CSR proxy score

of 21.192 with a standard deviation of 5.325. The highest mean value in the CSR proxy score

is obtained by Hungary with a value of 68.062 and a standard deviation of 19.130.

However, the number of observations for each country differs considerably. The sample

includes 71 companies from India representing 10.81% of the overall sample (Table 2), so that

it is relatively safe to attest a high mean firm valuation for Indian firms compared to other global

emerging markets. This might be due to the strong GDP growth rates the Indian economy

experienced in the past decade which are likely to drive future earnings growth and should

ultimately be reflected in market valuations. In the other cases, the final sample merely includes

five companies for Egypt and three companies for Hungary, both representing less than one

percent of the overall sample. Therefore, I refrain from drawing generalized conclusions about

firm valuation and CSR performance levels from these findings.

Instead, panel b reveals that when countries are combined into their respective MSCI regions,

the average values and standard deviations of both variables of EMEA and Americas are very

similar, so that the cross-country differences when aggregated might play less of a role.

Companies across emerging markets in Asia, however, seem to have higher average firm

valuations than their counterparts in EMEA and the Americas. Specifically, average TOBQ of

1.712 for Asian companies is higher than 1.578 for companies located in EMEA and 1.598 for

companies located in the Americas. At the same time, the standard deviation of TOBQ of 1.168

across Asian firms is considerably higher compared to the other two regions, with 0.992 for

EMEA and 0.833 for Americas. While Asian firms enjoy relatively large valuations, these

valuations also differ to the greatest extent from company to company. At the same time, they

exhibit the worst average CSR performance of 43.959, which is below the 49.660 of EMEA

and the 52.892 of Americas. The standard deviation in the CSR dimension is reasonably equal

across all three regions.

41

Table 4: TOBQ and CSR by country and region

This table presents the descriptive statistics for the key variables of the full unbalanced data panel covering 3,800

firm-year observations of 657 individual emerging market companies over the period from 2010-2016 used for

baseline regression (5) by country and region. Firm valuation is the dependent variable and proxied by one-year

ahead Tobin’s q, which is defined as 𝑀𝑎𝑟𝑘𝑒𝑡 𝐶𝑎𝑝𝑖𝑡𝑎𝑙𝑖𝑧𝑎𝑡𝑖𝑜𝑛𝑖,𝑌+𝐵𝑜𝑜𝑘 𝑉𝑎𝑙𝑢𝑒 𝑜𝑓 𝐴𝑠𝑠𝑒𝑡𝑠𝑖,𝑌−𝐵𝑜𝑜𝑘 𝑉𝑎𝑙𝑢𝑒 𝑜𝑓 𝐸𝑞𝑢𝑖𝑡𝑦𝑖,𝑌

𝐵𝑜𝑜𝑘 𝑉𝑎𝑙𝑢𝑒 𝑜𝑓 𝐴𝑠𝑠𝑒𝑡𝑠𝑖,𝑌 and

collected for the financial reporting years from 2011-2017. Corporate Social Responsibility (CSR) is the

independent variable and proxied by Thomson Reuters EIKON ESG scores as 𝐸𝑛𝑣𝑖𝑟𝑜𝑛𝑚𝑒𝑛𝑡𝑎𝑙 𝑝𝑖𝑙𝑙𝑎𝑟 𝑠𝑐𝑜𝑟𝑒𝑖,𝑌+𝑆𝑜𝑐𝑖𝑎𝑙 𝑝𝑖𝑙𝑙𝑎𝑟 𝑠𝑐𝑜𝑟𝑒𝑖,𝑌

2 and collected over the period from 2010-2016. Panel a shows the

mean and standard deviation of TOBQ and CSR in 20 different countries defined as emerging markets according

to the Morgan Stanley Capital International (MSCI) Emerging Market (EM) Index classification. Panel b shows

the mean and standard deviation of TOBQ and CSR across the three regions in which the MSCI EM index is

divided. The raw data have been retrieved for companies in the Thomson Reuters EIKON Global Emerging Market

Index which have their headquarters based in a country defined as emerging market by the Morgan Stanley Capital

International (MSCI) Emerging Market (EM) Index classification. Next to firms with insufficient data availability,

firms with negative Book value of Equity, TOBQ >8, and financial sector firms (GICS: Banks and Insurances)

have been excluded. The sample is further trimmed at the top and bottom percent across SIZE and ROA.

Panel A: Descriptive statistics by country

TOBQ CSR

Country Mean SD Mean SD

South Africa 1.881 1.161 56.199 15.408

Malaysia 1.656 0.827 46.384 15.073

Hong Kong 1.179 0.412 33.796 16.401

Taiwan 1.509 0.918 40.399 22.603

Thailand 2.158 1.456 58.909 18.610

Philippines 1.806 0.787 45.577 20.682

Indonesia 2.161 1.201 47.105 19.194

India 2.445 1.824 53.942 20.748

Czech Republic 0.996 0.141 39.780 5.211

Russia 1.223 0.699 45.926 19.274

Turkey 1.630 0.859 51.772 18.880

Hungary 1.135 0.255 68.062 19.130

Brazil 1.526 0.883 58.960 17.739

China 1.250 0.400 36.255 17.535

Mexico 1.919 0.836 48.711 22.199

Egypt 0.980 0.210 21.192 5.325

Chile 1.413 0.554 41.969 19.384

United Arab Emirates 1.412 0.355 40.276 14.097

Qatar 1.608 0.640 21.872 12.711

Poland 1.169 0.709 40.876 21.560

Panel B: Descriptive statistics by MSCI region

TOBQ CSR

Region Mean SD Mean SD

Asia 1.712 1.168 43.959 21.025

EMEA 1.578 0.992 49.660 19.556

Americas 1.598 0.833 52.892 20.473

42

4.2 Inferential statistics

4.2.1 The effect of CSR performance on firm value across global emerging markets

The first column of table 5 below presents the results of the baseline analysis, i.e. the effect of

CSR rating performance (CSR) on one-year ahead firm value (TOBQ). Columns (2)-(4) show

regional differences in the effect of the baseline regressions following the MSCI classified

regions Asia, EMEA, and Americas, respectively. The constant in all specifications is relatively

large, because the logarithmic values of total assets, i.e. SIZE, are large in absolute values

compared to the other variables and thus shift up the regression line minimizing the weighted

squared residuals. The adjusted R2 of the overall baseline model in column (1) amounts to 0.651,

so the model explains a considerable portion of the overall variability of the data. All

coefficients are statistically significant and untabulated results show that dropping control

variables does not significantly increase the adjusted R2.

Column (1) shows a positive CSR coefficient of 0.181, which is statistically significant at the

1% level. Thus, a one-standard-deviation increase in normalised CSR score performance

(0.233) of an emerging market company in this sample is – on average – associated with a 0.042

(0.181*0.233) point increase in one-year ahead TOBQ. Compared to the mean value of 1.661

for TOBQ across the sample, this increase constitutes around 2.5% (0.042/1.661) of that value.

As TOBQ measures the market value of a firm’s assets over by the replacement value of the

firm’s assets and I know that the mean and median book values of total assets for this sample

is around USD 5.3bn, it is obvious that small changes in this measure can have large absolute

economic impacts. Thus, the analysis attests both a statistically and economically significant

value-enhancing effect of CSR performance across global emerging market firms.

In line with my predictions and former research, profitability is positively related to TOBQ, and

leverage and size are negatively related to TOBQ. A one-standard-deviation increase in ROA

(0.073) is associated with a 0.416 (5.705*0.073) point increase in one-year ahead TOBQ. This

large magnitude is intuitive given the sensitivity of firm valuation to profitability. A one-

standard-deviation increase in LEV (0.16) is associated with a 0.029 (0.184*0.16) point

decrease in one-year ahead TOBQ. As LEV is also constructed as a simple ratio from the

reporting data, firm valuation is less sensitive to LEV compared to ROA. This finding is in line

with the study of El Ghoul et al. (2017), which find a strong effect for return on assets but only

mild evidence for a negative effect of leverage.

43

Table 5: Baseline and regional TOBQ regressions

The first column presents the baseline unbalanced panel regression (5) results of the effect of corporate social

performance on firm valuation for a sample of 657 individual emerging market firms with a total of 3,800 firm-

year observations. Firm valuation is the dependent variable and proxied by one-year ahead Tobin’s q, which is

defined as 𝑀𝑎𝑟𝑘𝑒𝑡 𝐶𝑎𝑝𝑖𝑡𝑎𝑙𝑖𝑧𝑎𝑡𝑖𝑜𝑛𝑖,𝑌+𝐵𝑜𝑜𝑘 𝑉𝑎𝑙𝑢𝑒 𝑜𝑓 𝐴𝑠𝑠𝑒𝑡𝑠𝑖,𝑌−𝐵𝑜𝑜𝑘 𝑉𝑎𝑙𝑢𝑒 𝑜𝑓 𝐸𝑞𝑢𝑖𝑡𝑦𝑖,𝑌

𝐵𝑜𝑜𝑘 𝑉𝑎𝑙𝑢𝑒 𝑜𝑓 𝐴𝑠𝑠𝑒𝑡𝑠𝑖,𝑌 and collected for the financial

reporting years from 2011-2017. Corporate Social Responsibility performance (CSR) is the independent variable

and proxied by Thomson Reuters EIKON ESG scores as 𝐸𝑛𝑣𝑖𝑟𝑜𝑛𝑚𝑒𝑛𝑡𝑎𝑙 𝑝𝑖𝑙𝑙𝑎𝑟 𝑠𝑐𝑜𝑟𝑒𝑖,𝑌+𝑆𝑜𝑐𝑖𝑎𝑙 𝑝𝑖𝑙𝑙𝑎𝑟 𝑠𝑐𝑜𝑟𝑒𝑖,𝑌

2 and

collected over the period from 2010-2016. The control variables comprise Return on Assets (ROA) defined as Net income before extraordinary items𝑖,𝑌

Total Book Value of Assets𝑖,𝑌−1, Leverage (LEV) defined as


Book Value of Assets𝑖,𝑌, SIZE defined as the

logarithm of 𝐵𝑜𝑜𝑘 𝑉𝑎𝑙𝑢𝑒 𝑜𝑓 𝐴𝑠𝑠𝑒𝑡𝑠𝑖,𝑌, GDP defined as the logarithm of

GDP per capita in constant 2010 USD𝑖,𝑌, and Age defined as the logarithm of current Year 𝑌𝑖 −𝑌𝑒𝑎𝑟 𝑜𝑓 𝑖𝑛𝑐𝑜𝑟𝑝𝑜𝑟𝑎𝑡𝑖𝑜𝑛𝑖 . Columns (2)-(4) test for differences in the baseline effect across the three regions in

which the 657 companies are grouped, by containing the interaction terms CSR*Asia, CSR*EMEA,

CSR*Americas, respectively. The models are specified with Error Generalized Least Squares (EGLS) estimation

with cross-section weights, allowing for heterogeneity in the cross-sections. Furthermore, the two-tailed p-values

are based on robust standard errors adjusted for the presence of heteroskedasticity in the cross-sections. * and **

indicate significance at the 5 percent and 1 percent levels, respectively. The raw data have been retrieved for

companies in the Thomson Reuters EIKON Global Emerging Market Index which have their headquarters based

in a country defined as emerging market by the Morgan Stanley Capital International (MSCI) Emerging Market

(EM) Index classification. Next to firms with insufficient data availability, firms with negative Book value of

Equity, TOBQ >8, and financial sector firms (GICS: Banks and Insurances) have been excluded. The sample is

further trimmed at the top and bottom percent across SIZE and ROA.

Variable (1) (2) (3) (4)

C 7.058** 6.883** 7.065** 7.295**

CSR 0.181** 0.169** 0.235** 0.104**

ROA 5.705** 5.654** 5.638** 5.810**

LEV -0.184** -0.177** -0.193** -0.189**

SIZE -0.206** -0.201** -0.211** -0.211**

GDP -0.173** -0.168** -0.160** -0.184**

AGE 0.076** 0.081** 0.073** 0.074**

CSR*Asia 0.012

CSR*EMEA -0.257**

CSR*Americas 0.275**

Cross-section weights YES YES YES YES

Adj. R2 0.651 0.630 0.624 0.679

As SIZE, GDP and AGE are measured on a logarithmic scale, I assess their impact in percentage

changes rather than standard deviations, as I find these more intuitive. A company with a one

percent increase in SIZE is on average associated with a 0.00206 (0.206/100) point decrease in

one-year ahead TOBQ. In contrast to El Ghoul et al. (2017), the effect of GDP is negative and

significant in the global emerging market sample at hand. A company in a country with a one

percent increase in GDP per capita is on average associated with a 0.00173 (0.173/100) point

decrease in one-year ahead TOBQ. Moreover, as expected, firm age has a positive and

significant effect on TOBQ. A company with an increase in firm age is on average associated

with a 0.00076 (0.076/100) point increase in one-year ahead TOBQ. This result is not trivial,

as it still amounts to 0.05% of the mean value of TOBQ.

44

Columns (2)-(4) inspect regional differences in the baseline analysis by adding the interaction

terms CSR*Asia, CSR*EMEA, and CSR*Americas, respectively. For Asian companies, which

represent 60% of the sample, the CSR coefficient of 0.169 only slightly differs from the baseline

coefficient of 0.181 and the coefficient of 0.012 on CSR*Asia is not significant. Thus, the

overall effect seems to accurately reflect the effect present among companies from Asian

emerging markets. For companies from EMEA, which represent 20% of the sample, the CSR

coefficient of 0.235 exceeds the baseline coefficient of 0.181 and the negative coefficient of -

0.257 on CSR*Asia is significant. Thus, the overall effect seems not to accurately reflect the

effect present among companies in EMEA, as their individual effect amounts to -0.022 (0.235-

0.257). This coefficient is very likely to be indistinguishable from zero, so the overall value-

enhancing effect of CSR is absent for companies in EMEA. Their presence drags down the

coefficient of the overall regression across all three regions. For companies from the Americas,

which represent the other 20% of the sample, the CSR coefficient of 0.104 is considerably

below the baseline coefficient of 0.181 and the positive coefficient of 0.275 on CSR*Americas

is significant. Thus, the overall effect seems not to accurately reflect the effect present among

companies in the Americas, as their individual effect amounts to 0.456 (0.181+0.275). Their

presence pushes up the coefficient of the overall regression across all three regions.

Table 6 below shows the attempt of gaining additional statistical confidence in the overall

value-enhancing effect obtained in the baseline regression. As this is an unbalanced data panel

with varying number of observations per period, I cannot apply the Fama and MacBeth (1973)

method of determining the simple average cross-sectional coefficient and calculate a time-series

standard error of this average as a robustness check employed in similar analyses (e.g. Gompers

et al., 2003). The problem of the time-varying number of observations would be eliminated in

a balanced data panel. However, as mentioned before, due to the limited data availability and

potential attrition bias, an unbalanced panel setup is preferable for the data at hand. While this

trade-off between data quality and statistical ease is certainly a limitation, I can still gain

additional confidence in the cross-sectional dimension of the baseline analysis by conducting

year-by-year linear OLS regressions with heteroskedasticity and autocorrelation (HAC)

adjusted standard errors. As Table 6 below shows, the coefficient of CSR is positive across all

seven financials years.

In five out of these seven years, the coefficients are also statistically significant. They vary from

0.134 to 0.395 and in six out of seven years, they are higher than the coefficient of 0.181 of the

baseline panel regression above. The adjusted R2 of the cross-sectional OLS regressions varies

from 0.412 to 0.495 and thus remains strictly below the value of 0.651 in our baseline panel

45

regression above. This result shows that bundling the cross-sectional and time-series

information in the data panel above increases the predictive ability of the model compared to

year-by-year cross-sectional regressions by around 30-60%. The effect of all other variables is

also in line with the baseline panel regression above. The only exception is leverage. While

LEV has negative coefficients across all years, they are – in contrast to the panel analysis – not

significant in any of the years.

Table 6: Year-by-year baseline TOBQ regressions

This table presents the year-by-year effects of corporate social performance on firm valuation for the period of

2010-2016. Firm valuation is the dependent variable and proxied by one-year ahead Tobin’s q, which is defined

as 𝑀𝑎𝑟𝑘𝑒𝑡 𝐶𝑎𝑝𝑖𝑡𝑎𝑙𝑖𝑧𝑎𝑡𝑖𝑜𝑛𝑖,𝑌+𝐵𝑜𝑜𝑘 𝑉𝑎𝑙𝑢𝑒 𝑜𝑓 𝐴𝑠𝑠𝑒𝑡𝑠𝑖,𝑌−𝐵𝑜𝑜𝑘 𝑉𝑎𝑙𝑢𝑒 𝑜𝑓 𝐸𝑞𝑢𝑖𝑡𝑦𝑖,𝑌

𝐵𝑜𝑜𝑘 𝑉𝑎𝑙𝑢𝑒 𝑜𝑓 𝐴𝑠𝑠𝑒𝑡𝑠𝑖,𝑌 and collected for the financial reporting years

from 2011-2017. Corporate Social Responsibility performance (CSR) is the independent variable and proxied by

Thomson Reuters EIKON ESG scores as 𝐸𝑛𝑣𝑖𝑟𝑜𝑛𝑚𝑒𝑛𝑡𝑎𝑙 𝑝𝑖𝑙𝑙𝑎𝑟 𝑠𝑐𝑜𝑟𝑒𝑖,𝑌+𝑆𝑜𝑐𝑖𝑎𝑙 𝑝𝑖𝑙𝑙𝑎𝑟 𝑠𝑐𝑜𝑟𝑒𝑖,𝑌

2 and collected over the

period from 2010-2016. Control variables comprise Return on Asset (ROA) defined as Net income before extraordinary items𝑖,𝑌




logarithm of 𝐵𝑜𝑜𝑘 𝑉𝑎𝑙𝑢𝑒 𝑜𝑓 𝐴𝑠𝑠𝑒𝑡𝑠𝑖,𝑌 , GDP defined as the logarithm of

GDP per capita in constant 2010 USD𝑖,𝑌, and Age defined as the logarithm of current Year 𝑌𝑖 −

𝑌𝑒𝑎𝑟 𝑜𝑓 𝑖𝑛𝑐𝑜𝑟𝑝𝑜𝑟𝑎𝑡𝑖𝑜𝑛𝑖 . The first two columns of this table present the financial years in which the data of the

independent variables have been collected with the respective number of observed firms. In the other columns, the

estimated ordinary least squares (OLS) regression coefficients of all independent variables with the corresponding

heteroskedasticity and autocorrelation adjusted (HAC) standard errors in brackets below are depicted. The last

column depicts the adjusted R2 of the respective regression per year. * and ** indicate significance at the 5 percent

and 1 percent levels, respectively. The raw data have been retrieved for companies in the Thomson Reuters EIKON

Global Emerging Market Index which have their headquarters based in a country defined as emerging market by

the Morgan Stanley Capital International (MSCI) Emerging Market (EM) Index classification. Next to firms with

insufficient data availability, firms with negative Book value of Equity, TOBQ >8, and financial sector firms

(GICS: Banks and Insurances) have been excluded. The sample is further trimmed at the top and bottom percent

across SIZE and ROA.

YEAR N C CSR ROA LEV SIZE GDP AGE Adjusted

R2

2010 424 7.954** 0.352** 4.793** -0.125 -0.270** -0.124** 0.0756 0.426

(0.865) (0.151) (1.126) (0.234) (0.035) (0.042) (0.050)

2011 470 9.638** 0.395* 5.763** -0.194 -0.332** -0.177** 0.164** 0.456

(1.084) (0.194) (1.068) (0.274) (0.041) (0.043) (0.059)

2012 517 8.400** 0.134 7.806** -0.146 -0.278** -0.168** 0.134** 0.485

(1.091) (0.162) (1.487) (0.241) (0.035) (0.051) (0.061)

2013 551 10.437** 0.294* 6.989** -0.601 -0.284** -0.324** 0.044 0.462

(1.438) (0.146) (1.857) (0.310) (0.039) (0.067) (0.062)

2014 595 8.036** 0.055 7.304** -0.172 -0.209** -0.250** 0.018 0.412

(1.094) (0.137) (0.791) (0.212) (0.036) (0.047) (0.064)

2015 618 7.530** 0.245* 8.824** -0.074 -0.198** -0.235** 0.021 0.495

(0.954) (0.119) (0.880) (0.168) (0.029) (0.036) (0.056)

2016 625 8.598** 0.320* 8.010** -0.334 -0.224** -0.285** 0.047 0.430

(1.031) (0.152) (0.993) (0.250) (0.031) (0.045) (0.073)

46

4.2.2 The role of firm-level governance and ESG rating categories

Table 7 below illustrates the results of several model specifications aimed at gaining a deeper

understanding of the role of firm-level governance performance and the accuracy of the ESG

rating methodology. First, I analyse models (1)-(3) to assess the role of firm-level governance

rating performance on firm value. Column (1) shows that the overall ESG combined score has,

just as the CSR proxy, a positive and significant effect on one-year ahead TOBQ. With a

coefficient of 0.153, the effect is slightly weaker compared to the coefficient of 0.181 of the

baseline data panel regression. A one-standard-deviation increase in normalised ESG combined

score performance (0.198) of an emerging market company in this sample is – on average –

associated with a 0.030 (0.153*0.198) point increase in one-year ahead TOBQ. Column (2)

shows that when disseminating the ESG scores into the three individual pillar scores, each pillar

yields positive and significant coefficients. The governmental pillar shows the largest

coefficient of 0.147, followed by the environmental pillar with a coefficient of 0.080 and the

social pillar with a coefficient of 0.065.

Column (3) shows a positive and significant coefficient of 0.141 for the governance pillar score

individually. A one-standard-deviation increase in normalised governance pillar score

performance (0.222) of an emerging market company in this sample is – on average – associated

with a 0.031 (0.141*0.222) point increase in one-year ahead TOBQ. Compared to the mean

value of 1.661 for TOBQ across the sample, this increase constitutes around 1.9% (0.031/1.661)

of that value and thus economically significant. Compared to the 2.5% economic impact of CSR

in the baseline regression, the effect of firm-level governance is weaker. Interestingly, the

governance pillar alone seems to explain more variability in TOBQ than the three pillars

together as well as the overall ESG combined score, as shown by the highest adjusted R2 of

0.644 in column (3). However, the differences in adjusted R2 amount only to around 5% and

are thus marginal.

Second, I analyse the methodology employed by Thomson Reuters in constructing their ESG

scores by dissecting the pillar scores into their category scores in columns (4) and (5). Column

(4) splits the governance pillar into its category scores: management, shareholders, and CSR

strategy. While the management and CSR strategy categories both show positive and significant

(1% level) coefficients, the shareholders category has a significant (5% level) negative effect

of -0.045 on firm value. The positive effect at pillar level seems to be driven by the dominance

of the positive effects of management and CSR strategy, but weakened by the shareholder

category. However, the adjusted R2 in column (4) is slightly lower than in column (3), meaning

that the splitting of the governance pillar does not increase the predictive power of the model.

47

In column (5), all category scores of the three pillars are included. In this specification, the

adjusted R2 increases by around 9% to 0.673 compared to the overall ESG combined score of

0.618 in column (1). The positive coefficients of management and CSR strategy remain

consistent and significant, while the coefficient of the shareholder category remains negative

with -0.025 and becomes insignificant. These results combined provide some evidence of a

negative effect of the shareholder score on firm value. At least, the analysis suggests that the

shareholders rating performance has no significant effect on firm valuation.

Furthermore, column (5) reveals several inconsistencies in the categories of the other pillars.

Notably, the emission score of the environmental pillar has a negative and significant

coefficient, while the resource use and environmental innovation scores have a positive and

significant coefficient. Moreover, within the social pillar, the human rights score has a negative

effect, which is significant at the 5% level. The community score is positive and significant at

the 1% level, whereas the product responsibility score is insignificant. These findings are in

line with the criticism of Utz (2017) and Attig et al. (2013) and suggests that the rating

methodology of Thomson Reuters might not be ideal in terms of accuracy and that the common

CSR proxy above is contestable.

48

Table 7: Firm-level governance and ESG rating categories TOBQ regressions

This table presents the unbalanced panel regression results of the effect of various Thomson Reuters EIKON ESG

rating scores on firm valuation for a sample of 657 individual emerging market firms with a total of 3,800 firm-

year observations. Firm valuation is the dependent variable and proxied by one-year ahead Tobin’s q (TOBQ),

which is defined as 𝑀𝑎𝑟𝑘𝑒𝑡 𝐶𝑎𝑝𝑖𝑡𝑎𝑙𝑖𝑧𝑎𝑡𝑖𝑜𝑛𝑖,𝑌+𝐵𝑜𝑜𝑘 𝑉𝑎𝑙𝑢𝑒 𝑜𝑓 𝐴𝑠𝑠𝑒𝑡𝑠𝑖,𝑌−𝐵𝑜𝑜𝑘 𝑉𝑎𝑙𝑢𝑒 𝑜𝑓 𝐸𝑞𝑢𝑖𝑡𝑦𝑖,𝑌


reporting years from 2011-2017. The Thomson Reuters ESG combined score (row 7), the three pillar level scores

Environmental, Social, and Governance (rows 8-10) as well as the ten category level scores Resource Use,

Emissions, Environmental Innovation, Workforce, Human Rights, Community, Product Responsibility,

Management, Shareholders, and CSR Strategy (rows 11-20) are the dependent variables and collected for the

financial reporting years from 2010-2016. All these scores that originally range from 0-100 and are assessed

relative to their industry peers, have been normalised to range from 0-1. A detailed overview of the methodology

of the Thomson Reuters EIKON ESG ratings can be found in Appendix B. Control variables comprise Return on

Asset (ROA) defined as Net income before extraordinary items𝑖,𝑌



Book Value of Assets𝑖,𝑌, SIZE defined as the logarithm of 𝐵𝑜𝑜𝑘 𝑉𝑎𝑙𝑢𝑒 𝑜𝑓 𝐴𝑠𝑠𝑒𝑡𝑠𝑖,𝑌 , GDP defined as the

logarithm of GDP per capita in constant 2010 USD𝑖,𝑌, and Age defined as the logarithm of current Year 𝑌𝑖 −

𝑌𝑒𝑎𝑟 𝑜𝑓 𝑖𝑛𝑐𝑜𝑟𝑝𝑜𝑟𝑎𝑡𝑖𝑜𝑛𝑖 . The models are specified with Error Generalized Least Squares (EGLS) estimation with

cross-section weights, allowing for heterogeneity in the cross-sections. Furthermore, the two-tailed p-values are

based on robust standard errors adjusted for the presence of heteroskedasticity in the cross-sections. * and **

indicate significance at the 5 percent and 1 percent levels, respectively. The raw data have been retrieved for

companies in the Thomson Reuters EIKON Global Emerging Market Index which have their headquarters based

in a country defined as emerging market by the Morgan Stanley Capital International (MSCI) Emerging Market

(EM) Index classification. Next to firms with insufficient data availability, firms with negative Book value of

Equity, TOBQ >8, and financial sector firms (GICS: Banks and Insurances) have been excluded. The sample is

further trimmed at the top and bottom percent across SIZE and ROA.

Variable (1) (2) (3) (4) (5)

C 6.730** 6.761** 6.822** 6.845** 6.833**

ROA 5.571** 5.584** 5.508** 5.528** 5.925**

LEV -0.204** -0.149** -0.150** -0.160** -0.064

SIZE -0.192** -0.199** -0.204** -0.199** -0.208**

GDP -0.173** -0.169** -0.165** -0.174** -0.161**

AGE 0.089** 0.089** 0.112** 0.093** 0.081**

ESG combined score 0.153**

Environmental Pillar (E) 0.080**

Social Pillar (S) 0.065*

Governance Pillar (G) 0.147** 0.141**

Resource Use 0.217** Emissions -0.228** Environmental Innovation 0.203**

Workforce -0.013 Human Rights -0.067* Community 0.137** Product Responsibility -0.013 Management 0.152** 0.119**

Shareholders -0.045* -0.025

CSR Strategy 0.077** 0.049**

Cross-section weights YES YES YES YES YES

Adj. R2 0.618 0.621 0.644 0.620 0.673

49

4.2.3 The role of sell-side analysts on CSR and firm value

The results of Table 8 below, relating to the role of sell-side analysts in the context of CSR and

firm value, are based on the smaller sample size of 638 individual firms with 3504 firm-year

observations. Column (1) repeats the baseline regression applied to this smaller sample. The

CSR performance coefficient is positive and significant with 0.142, which is slightly below the

0.181 found for the full sample.

In column (2), CSR is dropped, and ANA is added instead. The coefficient of 0.016 on ANA is

significant at the 1% level. The adjusted R2 increases from 0.632 to 0.663. Ignoring CSR, a

company with a one percent increase in the number of sell-side analysts is – on average –

associated with a 0.00156 (0.156/100) point increase in one-year ahead TOBQ. In column (3),

ANA is added next to CSR to the baseline model in column (1). Compared to model (1), the

coefficient of CSR drops from 0.142 to 0.104 and remains significant. Compared to model (2),

the coefficient of ANA only decreases marginally from 0.016 to 0.015. The adjusted R2

increases compared to the baseline model from 0.632 to 0.654 but decreases compared to

column (2) from 0.663 to 0.654. All in all, the value-enhancing effect of analyst coverage for

global emerging market companies seems to be robust.

In column (4), I include an additional interaction term CSR*ANA to the model of column (3)

to inspect whether the number of analysts reinforces or mitigates the positive effect of CSR on

firm value. Both the CSR coefficient of 0.223 and the ANA coefficient of 0.020 show the

strongest individual positive effect compared to the other specifications (1)-(3). The interaction

coefficient on CSR*ANA loads negatively with -0.009 and is significant at the 5% level. Thus,

the number of analysts has a mitigating effect on the positive relation between CSR

performance and firm valuation, supporting the view that both analysts as external monitors in

economies with weak institutions and CSR as internal strategic response to overcome

institutional voids are partly substitutes for each other. Specifically, a 100 percent increase in

the number of sell-side analysts covering an emerging market company in this sample on

average decreases the positive effect of CSR on one-year TOBQ by 0.009 from 0.223 to 0.194.

In this specification, the adjusted R2 increases to its highest value of 0.688 within the whole

inferential analysis. However, the models of the baseline analysis and the role of firm-level

governance are not directly comparable in terms of adjusted R2, because they are based on a

slightly larger sample. In summary, both CSR performance and the number of analysts

individually and together are associated with higher valuations and ANA has a mitigating effect

on the positive relation between CSR and firm value.

50

Table 8: Analyst TOBQ regressions

This table presents the additional analysis on the role of analyst coverage on the effect of corporate social

performance on firm valuation for a reduced sample of 638 individual emerging market firms with a total of 3,504

firm-year observations. Firm valuation is the dependent variable and proxied by one-year ahead Tobin’s q, which

is defined as 𝑀𝑎𝑟𝑘𝑒𝑡 𝐶𝑎𝑝𝑖𝑡𝑎𝑙𝑖𝑧𝑎𝑡𝑖𝑜𝑛𝑖,𝑌+𝐵𝑜𝑜𝑘 𝑉𝑎𝑙𝑢𝑒 𝑜𝑓 𝐴𝑠𝑠𝑒𝑡𝑠𝑖,𝑌−𝐵𝑜𝑜𝑘 𝑉𝑎𝑙𝑢𝑒 𝑜𝑓 𝐸𝑞𝑢𝑖𝑡𝑦𝑖,𝑌


reporting years from 2011-2017. Corporate Social Responsibility performance (CSR) is the independent variable

and proxied by Thomson Reuters EIKON ESG scores as 𝐸𝑛𝑣𝑖𝑟𝑜𝑛𝑚𝑒𝑛𝑡𝑎𝑙 𝑝𝑖𝑙𝑙𝑎𝑟 𝑠𝑐𝑜𝑟𝑒𝑖,𝑌+𝑆𝑜𝑐𝑖𝑎𝑙 𝑝𝑖𝑙𝑙𝑎𝑟 𝑠𝑐𝑜𝑟𝑒𝑖,𝑌

2 and

collected over the period from 2010-2016. Control variables comprise Return on Asset (ROA) defined as Net income before extraordinary items𝑖,𝑌




logarithm of 𝐵𝑜𝑜𝑘 𝑉𝑎𝑙𝑢𝑒 𝑜𝑓 𝐴𝑠𝑠𝑒𝑡𝑠𝑖,𝑌 , GDP defined as the logarithm of

GDP per capita in constant 2010 USD𝑖,𝑌, and Age defined as the logarithm of current Year 𝑌𝑖 −𝑌𝑒𝑎𝑟 𝑜𝑓 𝑖𝑛𝑐𝑜𝑟𝑝𝑜𝑟𝑎𝑡𝑖𝑜𝑛𝑖 . Analyst coverage (ANA) is defined as the logarithm of the 𝑊𝑒𝑖𝑔ℎ𝑡𝑒𝑑 −𝑎𝑣𝑒𝑟𝑎𝑔𝑒 𝑛𝑢𝑚𝑏𝑒𝑟 𝑜𝑓 𝑠𝑒𝑙𝑙 − 𝑠𝑖𝑑𝑒 𝑎𝑛𝑎𝑙𝑦𝑠𝑡𝑠𝑖,𝑌 that is covering the stock of the company in the respective financial

years 2010-2016. Column (1) tests the baseline regression model (5) for this reduced sample. Columns (2)-(3) test

the effect of ANA on TOBQ independently of and next to CSR. Column (4) tests the effect of ANA on the effect

of CSR on TOBQ by containing an interaction term CSR*ANA. The models are specified with Error Generalized

Least Squares (EGLS) estimation with cross-section weights, allowing for heterogeneity in the cross-sections.

Furthermore, the two-tailed p-values are based on robust standard errors adjusted for the presence of

heteroskedasticity in the cross-sections. * and ** indicate significance at the 5 percent and 1 percent levels,

respectively. The raw data have been retrieved for companies in the Thomson Reuters EIKON Global Emerging

Market Index which have their headquarters based in a country defined as emerging market by the Morgan Stanley

Capital International (MSCI) Emerging Market (EM) Index classification. Next to firms with insufficient data

availability, firms with negative Book value of Equity, TOBQ >8, and financial sector firms (GICS: Banks and

Insurances) have been excluded. The sample is further trimmed at the top and bottom percent across SIZE and

ROA.

Variable (1) (2) (3) (4)

C 7.060** 6.991** 6.912** 6.912** CSR 0.142** 0.104** 0.223** ROA 5.812** 5.642** 5.659** 5.773** LEV -0.171** -0.075** -0.054 -0.013 SIZE -0.212** -0.239** -0.240** -0.241** GDP -0.152** -0.117** -0.106** -0.114** AGE 0.063** 0.113** 0.107** 0.109** ANA 0.016** 0.015** 0.020**

CSR*ANA -0.009*

Cross-section weights YES YES YES YES

Adj. R2 0.632 0.663 0.654 0.688

51

5. Discussion & Limitations

First, the discussion chapter uses the statistical results described in the previous section to

confirm or to reject the hypotheses developed in the literature review and provides a subjective

interpretation of the results in the context of the discussed literature. This is done in a

chronological structure from hypothesis one to four. Second, a critical review of the methods

and concepts applied in the analysis, together with their limitations, are presented.

H1: CSR performance is positively related to firm value across global emerging markets

The baseline regression (5) in the first column of table 5 shows a positive and significant

coefficient for the CSR proxy. This effect is also evident in the reduced sample for the additional

analysis on the role of analysts illustrated in the first column of table 8. Furthermore, the year-

by-year cross-sectional regressions with OLS estimators and HAC robust standard errors of

table 6 show positive coefficients in all seven years, where five of them are statistically

significant. Therefore, the positive effect of hypothesis 1 is confirmed across two samples and

methodologies. In this sample of global emerging market companies, CSR performance is

positively related to (subsequent) firm valuation.

However, columns (2)-(4) of table 5 reveal that this overall result masks some important

regional differences. Specifically, while the effect of Asian companies is in line with the overall

results, there seems to be no effect or even a negative effect between CSR and TOBQ for

companies in the EMEA region and a pronounced positive effect for companies in the Americas

region. This result has far-reaching consequences for the generalization of the results. First, it

casts doubt on the practice of treating global emerging market companies as a group due to their

institutional differences. Second, it illustrates the ESG data problem in emerging markets.

While it would be interesting to further inspect the factors driving the diverging results in

EMEA and the Americas, both regions individually represent only 20% of the already relatively

small sample. Thus, conducting separate empirical analyses with high statistical power is

difficult.

My main finding of the baseline regression provides empirical evidence for and is consistent

with the resource-based view of CSR and is the first to inspect the CSP and CFP debate across

global emerging markets. Emerging markets are generally characterized by weaker market

supporting institutions supporting economic exchanges. This often results in governance

52

concerns and ultimately higher risk of expropriation of minority shareholders. Consequently,

economic exchanges undertaken of and with these companies endure higher transaction costs

and they suffer from worse access to resources. To prevent undue transaction costs, emerging

market firms must take strategic actions to mitigate these concerns. My analysis shows that

acting responsibly is such a strategic response, as emerging market companies across the globe

with superior CSR performance enjoy higher firm valuations. As such, it lends support to the

recent findings of El Ghoul et al. (2017), that CSR initiatives reduce transaction costs and

improve access to resources in countries with weak institutions, which increases firm value. At

the same time, regional differences persist, and the above dynamics probably hold for emerging

market companies across Asia and the Americas but might not hold for companies in the EMEA

region.

In line with El Ghoul et al. (2017), I find a positive relationship between profitability and firm

value. In fact, there is strong empirical evidence for this relationship, as ROA yields a positive

and significant coefficient across the entire baseline and additional analyses. As expected, firms

that are more profitable have more scope to pay dividends, invest in necessary R&D projects to

ensure future growth and to sustain adverse market movements, so they enjoy higher valuations.

Further in line with El Ghoul et al. (2017), I find milder evidence for a negative relationship

between leverage and subsequent firm valuation. LEV shows a negative and significant

coefficient in almost all panel regression specifications across the baseline and additional

analyses. However, although coefficients of the year-by-year regressions in table 6 are negative

in all years, they are not significant. All in all, the analyses support the rationale that an increase

in leverage limits the flexibility of management decisions and renders firms to be more

vulnerable to market shocks, resulting in lower valuation.

In line with El Ghoul et al. (2017), I find strong evidence that smaller emerging market

companies have higher firm valuations. Just like for ROA, this relationship is significant across

the entire baseline and additional analyses. As expected, larger firms have more limited

investment opportunities and lower future growth prospects. Lower expected future cash flows

translate into lower firm valuations. In line with Gompers et al. (2003), I find relatively strong

empirical evidence for a positive effect of firm age on firm value in this emerging market

sample. AGE shows a positive and significant coefficient in all panel regressions and in two

out of seven year-by-year OLS regressions. As expected, more established firms in emerging

markets can draw on a larger history of treating stakeholders fairly and thus mitigate concerns

about expropriation, which results in higher firm valuation.

53

In contrast to El Ghoul et al. (2017), I find strong evidence for a negative relationship between

economic development proxied by the logarithm of GDP per capita and TOBQ for my sample

of exclusively emerging market companies. To illustrate, the observation with the lowest GDP

per capita belongs to India with USD 1,344, while the observation with the highest GDP per

capita belongs to Qatar with USD 70,306 (both in constant 2010 USD). While both are being

classified as emerging markets according to the MSCI EM index scheme used, the average

person in Qatar produces more than 50 times as much GDP as an average person in India. This

illustrates the wide range of economic development across the countries MSCI defines as

emerging. At the same time, GDP grew by 46% during the sampling period in India and fell by

7% in Qatar. The emerging countries at the bottom of the economic development have been

shown to yield higher growth potential than almost fully developed countries. Strong income

growth drives domestic demand and thus expected growth of domestic companies, which

results in higher expected future cash flows and is ultimately reflected in higher firm valuations.

Therefore, the wide dispersion in development levels in the MSCI classified emerging markets

with their different associated growth prospects provide rationale for this negative effect.

H2: Firm-level governance rating performance has a positive effect on firm value across global

emerging markets

In line with the theoretical suggestion of Ding et al. (2010) that firm-level governance is

especially important in emerging markets plagued by institutional voids and the empirical

finding of Kim et al. (2014) that in the U.S., governance ratings are associated with lower firm

risk, I find that across emerging market companies, governance rating performance is positively

associated with subsequent firm valuation. The governance pillar in table 7 shows a positive

and significant coefficient when regressed with the other two pillars in column (2) as well as

when regressed separately in column (3). Thus, the analysis confirms hypothesis 2.

At the same time, I find some evidence that the category scores incorporated in the governance

pillar are based on the traditional developed markets governance model dominated by PA

conflicts (Jensen and Meckling, 1976), but fail to properly incorporate the institutional

environment of emerging markets which results in the dominance of PP conflicts (Young et al.,

2008). When regressing the separate categories of the governance pillar on TOBQ in column

(4) of table 7, the shareholders score has a significant negative coefficient, but the adjusted R2

decreases. When regressing all category scores on TOBQ in column (5), this coefficient is still

negative, but insignificant. These results favour the view that unlike attested for developed

54

markets, more power of shareholders vs. manager across global emerging markets is not

associated with better firm performance, or even detrimental to firm performance (e.g. Faccio

et al., 2001). As explained above, emerging market companies are often dominated by majority

shareholders belonging to influential families or the state. In that context, even higher

shareholder power could for example mean that these majority shareholders steer business

according to their personal interests and might put affiliated state officials or family members

in the board of directors, rather than acting in the interest of all shareholders. This is likely to

lead to poor strategic choices or expropriation of minority shareholders and ultimately results

in lower firm valuations.

Columns (3)-(4) of table 7 show that the CSR strategy score, which TR includes in the

governance pillar and which is therefore not included in my CSR proxy, is positive and

significant. As explained before, CSR strategy refers for example to the establishment of a

sustainability committee, voluntarily disclosures, or sustainability audits. As such, a positive

and significant effect of this category score might be regarded as additional confirmatory

evidence for hypothesis 1. However, it also gives rise to criticism regarding the common

practice in academic studies of using the average of the environmental and social pillars as CSR

proxy, as this category score is part of the governance pillar but seems to capture important

parts of CSR activity.

Furthermore, column (4) of table 7 provides evidence in line with the criticism of Utz et al.

(2017) and Attig et al. (2013), that some category scores yield contradictory signs when

compared to their pillar scores. While there is a theoretical rationale for the negative effect of

the shareholder score on firm value in emerging markets, I find no such explanation for the

negative and significant effects of the resource use score and the human rights score on firm

valuation. Simultaneously, disaggregating the scores into all categories yields additional

predictive power, as the adjusted R2 of 0.673 is the highest across the different model

specifications. Thus, the inconsistent category scores can explain more variation in firm value

than the pillar scores and the ESG combined score. Thus, practitioners and academics need to

be careful when using TR ESG scores in their investment decisions. The CSR proxy based on

the TR methodology might not be ideal as the environmental and social pillars are not capturing

important parts of CSR activity contained in the governmental category CSR strategy.

Furthermore, lower-level category scores are in some cases inconsistent with the pillar scores.

55

H3: Analyst coverage has a positive effect on firm value across global emerging markets

Table 8 provides strong evidence for a positive effect of the number of sell-side analysts on

firm valuation across global emerging markets and thus confirms hypothesis 3. This positive

effect is consistent across all three model specifications in columns (2)-(4). It only varies

marginally between 0.015 and 0.020 and is significant at the 1% level. In line with Chung and

Jo (1996) and Yu (2008), this result suggests that equity analysts act as an external monitor and

help reducing agency costs, disciplining managers and steering investor attention towards

important information. The results contrast the view of Chan and Hameed (2006), which states

that analysts fail to produce firm-specific information in emerging economies that is being

valued by the market.

H4: Analyst coverage amplifies the positive effect of CSR on firm value across global emerging

markets

Column (4) of table 8 shows a negative coefficient of -0.009 on the interaction term of

CSR*ANA and adjusted R2 increases to 0.688. This coefficient is significant at the 5% level,

so I reject hypothesis 4. Instead, analyst coverage mitigates the effect of CSR on firm value

across global emerging markets. This mitigating effect can be explained by the rationale of Jo

and Harjoto (2014), who state that while analysts are primarily concerned with financial

information, they provide indirect but additional social pressure on firms to reduce their

irresponsible activities. While their analysis was confined to a sample of U.S. firms, this seems

to hold true across global emerging markets as well. The results suggest that analysts perform

to a certain degree the same task as ESG rating agencies, i.e. signalling trust to the marketplace

that the covered company is well governed and bears little risk of minority shareholder

expropriation.

At the same time, the individual positive effects of CSR and ANA on firm value found in

column (3) of table 8 remain significant when including the interaction term CSR*ANA in

column (4). Thus, despite ANA mitigating the positive effect of CSR on firm value, both

information intermediaries individually add value to the firm. This is supportive for the claim

of Berk and DeMarzo (2011), that the primary role of equity analysts is to uncover any financial

reporting irregularities, rather than providing elaborate ESG information. Consequently, while

analysts seem to provide some additional social pressure on firms to reduce their irresponsible

56

activities and signal trust in the proper governance of the firm, they rather complement ESG

rating agencies by focussing on financial reporting irregularities.

There are several limitations I acknowledge regarding my empirical analysis. First, my

classification of emerging market countries based on the MSCI EM index methodology is very

broad. As mentioned earlier, economic development, when measured as GDP per capita in

constant 2010 USD ranges from USD 1,344 in India to more than 50 times that value, i.e. USD

70,306, in Qatar. Despite being widely used in academic research, this classification scheme is

contestable and future research could define emerging markets more narrowly to produce

additional confidence in the results obtained. Furthermore, the descriptive analysis in table 3

shows that the companies covered do have an average and median book value of assets of

around USD 5.3bn, have been existing on average for 35 years, and are being followed – on

average – by 15 sell-side analysts. Thus, my results could be limited to relatively large and

mature companies rather than to hold for the entire spectrum of emerging market firms and

should therefore be treated with caution.

Furthermore, the inferential analysis of table 5 shows that the overall positive effect of CSR

performance on firm value of the baseline regression is driven by the dominant 60% share of

Asian companies in the sample. The effect seems to be considerably stronger for companies in

the Americas and non-existent for companies in EMEA. As such, the treatment of global

emerging markets as a group is contestable and further research on the underlying reasons for

the regional differences is needed. Moreover, I find and recognise drawbacks of using the

common CSR proxy of taking the average of the environmental and social pillar scores in table

6. Nevertheless, I rely on this proxy in the other analyses. Therefore, my results only hold to

the extent to which this proxy really does represent CSR performance. Besides, as mentioned

before, there is a lack of data on institutional shareholdings. Consequently, this analysis misses

to inspect the role of this important external governance mechanism.

Furthermore, a potentially endogenous relationship between CSR and TOBQ is a concern in

my analysis. Endogeneity broadly refers to situations in which an explanatory variable is

correlated with the error term, which can inflict bias in regression estimates and is mostly

caused by omitted variables, reverse causality, or simultaneity. Specifically, endogeneity would

arise due to unobservable heterogeneity from omitted firm-specific variables that are correlated

with CSR and TOBQ. Endogeneity from reverse causality would arise when superior financial

performance causes firms to improve their CSR performance, rather than superior CSR

57

performance resulting in higher firm valuation. Endogeneity from simultaneity would arise

when CSR and TOBQ are jointly determined, i.e. CSR performance and firm value

simultaneously affect each other (Dimson et al., 2015).

My research design helps to mitigate concerns about omitted firm-level heterogeneity by

including numerous control variables identified in similar previous studies that tested for

endogeneity. Nevertheless, it is possible that omitted country-level factors affect both CSR and

TOBQ and drive the results. More advanced econometrical methods of recent studies (e.g., El

Ghoul et al., 2011; Kim et al., 2014; El Ghoul et al., 2017), such as instrumental variable

techniques or dynamic panel generalized methods of moments (GMM) estimations could be

employed to further rule out omitted variable concerns. However, employing these methods

extends the scope of this thesis and I recognise that as a limitation of my analysis. The same

omitted variable problem also applies to the additional analyses regarding the relationship of

the firm-level governance and TOBQ as well as ANA and TOBQ.

To address the reverse causality and simultaneity problems, I follow previous research (e.g.

Kim et al., 2014; El Ghoul et al., 2017) and use a one-period lag between the dependent

variables including the CSR performance proxy and the independent variable TOBQ.

Nevertheless, this practice is imperfect since CSR scores are quite sticky across years. Besides,

while the EGLS specification of my unbalanced panel regression solves the problem of

heteroskedasticity in the remainder error term of the cross-section fixed effect specification and

reduces the Jarque-Bera statistic, the normality assumption of the residuals is still slightly

violated.

58

6. Conclusion

The conclusion chapter is structured as follows. First, it synthesizes the context of the academic

debate around which the analysis revolves and points out how the findings of the main and

additional analyses contribute to the respective existing literature streams and what implications

they have on future academic research. Second, recommendations for practitioners are derived.

Third, motivations for future research are presented. Limitations have already been stated on

the previous two pages and are therefore not repeated in this chapter.

While there is a large body of research on the CSP and CFP debate in developed markets

producing mixed results, research on the CSP and CFP debate in emerging markets is scarce.

So far, recent studies merely relate to stock price crash risk and are confined to individual

emerging markets (Zhang, Xie, and Xu, 2016; Lee, 2016). CSR research on emerging markets

as a group is absent. This is probably due to the failure to recognize institutional commonalities

across global emerging market companies, ESG data scarcity in emerging markets and the

extremely volatile capital markets in the last decade steering the focus on risk characteristics.

From an academic perspective, the main analysis of my thesis contributes to the current stance

of research by revitalizing the traditional deadlocked CSP and CFP debate in developed markets

by focussing on emerging markets and on the so far untouched post-financial crisis period of

2010-2016.

The rationale to conduct research across global emerging markets as a group is an extension of

the link between institutional theory, transaction cost theory and the CSP and CFP debate

recently found by El Ghoul et al. (2017). The absence of market-supporting institutions in an

economy often results in governance concerns and ultimately higher risk of expropriation of

minority shareholders of companies in that economy. Consequently, economic exchanges

undertaken of and with these companies endure higher transaction costs and they suffer from

worse access to resources. El Ghoul et al. (2017) claims that the traditional performance

channels proposed by the resource-based view of CSR like superior management

incentivization, moral capital, information quality, transparency, and trust, are expected to work

particularly well in the presence of institutional voids. They find that CSR constitutes a strategic

response to reduce the undue transaction costs associated with the absence of institutions and

creates necessary resources which ultimately result in higher firm valuation. As global

emerging markets as a group generally have relatively weak institutions (e.g. Meyer et al.,

59

2009), I posit that CSR performance across global emerging market companies is positively

related to firm valuation.

Indeed, I find that CSR performance proxied by the average of the environmental and social

pillar scores of the Thomson Reuters EIKON ESG rating database positively relates to firm

valuation proxied by one-year ahead Tobin’s q (TOBQ). Specifically, a one-standard-deviation

increase in normalised CSR score performance of an emerging market company in this sample

is – on average – associated with a 0.042-point increase in one-year ahead TOBQ. However,

this overall result is largely driven by the Asian companies representing around 60% of the

sample. Companies located in EMEA do not show any value enhancing effect of CSR and

companies located in the Americas show an especially strong value enhancing effect of CSR.

My main contribution is to enhance the current stance of research by revitalizing the deadlocked

CSP and CFP debate in developed markets. First, my study is the first study to provide empirical

evidence for the resource-based view of CSR in the so far untouched post-financial crisis period

of 2010-2016. Second, it is the first study that examines the link between CSR performance and

firm value across global emerging markets. It implies that future research should recognize that

in contrast to earlier suggestions (e.g. Baughn et al., 2007), emerging economies are sensitive

about CSR issues and that CSR performance is value-enhancing in those markets. At the same

time, it shows that large regional differences in this link do exist and that there is a need for

more extensive ESG data to conduct meaningful statistical analyses on regional level.

Furthermore, it provides confirmatory evidence for the proposed link between institutions and

transaction costs proposed by El Ghoul et al. (2017) and implies that future research on CSR

should recognize that the effectiveness of CSR performance channels depends on the

institutional context of the economy in which a company is located in.

The additional analysis on firm-level governance scores contributes to the corporate governance

and institutional theory literature. In line with suggestions that the need for effective internal

corporate governance is especially high in countries where institutional voids inhibit market

oversight or external governance (Ding et al., 2010), I find a positive effect of firm-level

governance scores on firm valuation. At the same time, further empirical tests suggest that

Thomson Reuters ESG scores are based on the traditional developed markets governance model

dominated by PA conflicts (Jensen and Meckling, 1976), but fail to properly incorporate the

institutional environment of emerging markets which results in the dominance of PP conflicts

(Young et al., 2008). The negative effect of the shareholders category score supports the view

60

of Faccio et al. (2001) that power towards shareholders vs. managers in emerging markets

increases PP conflicts instead of solving PA conflicts.

It implies for future research on corporate governance, that institutions, amongst other factors,

effect the governance requirements of companies and that these dynamics potentially prevent

generic governance scores from being universally applicable across different countries and

regions. Furthermore, the analysis finds further evidence in line with the criticism of Utz et al.

(2017) and Attig et al. (2013), that some category scores yield contradictory signs compared to

their pillar scores. For studies in both developed and emerging markets, academics should

recognize that the commonly used CSR proxy of taking the average of the environmental and

social pillar scores does not incorporate the CSR strategy component attributed to the

governance pillar even though it should conceptually include it, and that analyses on category

score level might produce misleading results.

Finally, this thesis contributes to the CSP and CFP debate by introducing another previously

untouched aspect which relates to the information intermediary role of ESG rating agencies,

i.e. analyst coverage. In line with Chung and Jo (1996) and Yu (2008), who suggest that equity

analysts act as an external monitor and help reducing agency costs, disciplining managers, and

steering investor attention towards important (financial) information, I find strong evidence for

a positive effect of the number of sell-side analysts on firm valuation (next to CSR). I further

find empirical support for a mitigating impact of analyst coverage on the positive effect of CSR

and firm value, while the individual coefficients remain positive and significant. This implies

that while analysts seem to provide some additional social pressure on firms to reduce their

irresponsible activities and signal trust in the proper governance of the firm, they rather

complement than substitute ESG rating agencies as information intermediaries by focussing on

financial reporting irregularities. It implies that future research on financial performance or

transaction costs in emerging markets should recognize that both analyst coverage and CSR

performance help reduce transaction costs and create resources that are valued by the market.

From a practitioner’s perspective, my findings have several implications on portfolio managers

of institutional investors and managers of emerging market companies. When allocating capital

to equities of emerging market companies, portfolio managers should pick stocks of companies

with strong future CSR capabilities, large analyst coverage, and conduct firm-level governance

analyses considering PP problems rather than (relying on generic governance scores based

primarily on) PA problems. Alternatively, the findings could be an artefact of portfolio

managers already incorporating these factors in their stock picking process. If they provide

61

immense amounts of capital to socially responsible emerging market companies that are largely

covered by analysts, they drive the market values of these companies, which is reflected in

Tobin’s q. Moreover, the analysis implies that practitioners who rely on Thomson Reuters

EIKON ESG scores in their investment decisions should use them with caution, as governance

scores based on a developed corporate governance model might be inaccurate for assessing the

corporate governance quality of emerging market companies.

For managers of emerging market companies, the analysis is relevant for their strategic

decision-making process, growth strategies, and capital budget allocations. My support for the

resource-based view of CSR calls on managers in emerging markets to strive for more CSR

excellence, which necessitates a strategic planning process. Managers of emerging market firms

seeking capital to finance identified growth opportunities might be well advised to increase

performance and communication of their CSR commitments to attract the desired capital.

Finally, the results immediately effect the capital budgeting process for managers deciding how

much resources to commit to CSR related activities.

Beyond tackling the issues immediately addressing the limitations presented in the previous

section (especially endogeneity), this study opens the gateway to a broad future research area.

As mentioned before, it is a pioneering study in inspecting the CSP and CFP debate in the

context of global emerging markets. Compared to previous research using data panels, the

sample construction of an unbalanced data panel with as much as 3,800 firm-year observations

shows that the data availability reached a tipping point where it is possible to conduct

meaningful analyses. Thus, numerous previous studies on CSR could be inspected in the

context of global emerging markets. One recent example would be to inspect the relationship

between CSR and stock price crash risk for this sample.

Besides, further research could be dedicated to an elaborate investigation of the regional

differences between the value relevance of CSR that have been revealed in this study. However,

more data would be needed than is currently accessible through Thomson Reuters EIKON.

Furthermore, future research could be conducted to find better proxies for CSR by for example

trying to combine the CSR strategy component score of the governmental pillar with the

environmental and social pillars. Alternatively, there might be opportunities to construct better

CSR proxies from different databases. On top of that, future research could be conducted on

institutional shareholdings as another external monitoring mechanism across global emerging

markets.

62

References

Attig, N., Boubakri, N., El Ghoul, S., & Guedhami, O. (2016). Firm internationalization and

corporate social responsibility. Journal of Business Ethics, 134(2), 171-197.

Attig, N., Cleary, S. W., El Ghoul, S., & Guedhami, O. (2014). Corporate legitimacy and

investment—Cash flow sensitivity. Journal of Business Ethics, 121, 297–314.

Attig, N., Ghoul, S. E., Guedhami, O., & Suh, J. (2013). Corporate social responsibility and

credit ratings. Journal of Business Ethics, 117, 679–694.

Arya, B., & Zhang, G. (2009). Institutional reforms and investor reactions to CSR

announcements: Evidence from an emerging economy. Journal of Management Studies,

46(7), 1089-1112.

Bae, K., Lim, C., Wei, J. (2006). Corporate governance and conditional skewness in the world’s

stock markets. Journal of Business 79, 2999–3028.

Baltagi, B. (2005). Econometric analysis of panel data (3rd edition). New Jersey: John Wiley

& Sons.

Barnea, A., & Rubin, A. (2010). Corporate social responsibility as a conflict between

shareholders. Journal of Business Ethics, 97, 71–86.

Baughn, C., Bodie, N., & McIntosh, J. (2007). Corporate social and environmental

responsibility in Asian countries and other geographical regions. Corporate Social

Responsibility and Environmental Management, 14(4), 189–205.

Bebchuk, L., Cohen, A. (2005). The costs of entrenched boards. Journal of Financial

Economics 78, 409–433.

Bénabou, R., & Tirole, J. (2010). Individual and corporate social responsibility. Economica,

77(305), 1-19.

Berk, J. and DeMarzo, P. (2011). Corporate Finance (2nd edition). Boston, MA: Prentice Hall.

Bogdanich, W. (2008). Heparin is now suspected in 62 fatalities across US. New York Times.

Brammer, S., Brooks, C., & Pavelin, S. (2006). Corporate social performance and stock returns:

UK evidence from disaggregate measures. Financial management, 35(3), 97-116.

Breusch, T. S., & Pagan, A. R. (1979). A simple test for heteroscedasticity and random

coefficient variation. Econometrica, 1287-1294.

Bushee, B. J., & Noe, C. F. (2000). Corporate disclosure practices, institutional investors, and

stock return volatility. Journal of Accounting Research, 171-202.

Callen, J. L., & Fang, X. (2013). Institutional investor stability and crash risk: Monitoring

versus short-termism? Journal of Banking & Finance, 37(8), 3047–3063.

Chahal, H., & Sharma, R. D. (2006). Implications of Corporate Social Responsibility on

Marketing Performance: A Conceptual Framework. Journal of Services Research, 6(1).

Chan, K., & Hameed, A. (2006). Stock price synchronicity and analyst coverage in emerging

markets. Journal of Financial Economics, 80(1), 115-147.

Chen, J., Hong, H., & Stein, J. C. (2001). Forecasting crashes: Trading volume, past returns,

and conditional skewness in stock prices. Journal of Financial Economics, 61(3), 345-

381.

63

Chung, K. and Jo, H. (1996). The impact of security analysts’ monitoring and marketing

functions on the market value of firms. Journal of Financial and Quantitative Analysis,

31(4), 493–512.

Claessens, S., Djankov, S. and Lang, L. (2000). The separation of ownership and control in East

Asian corporations. Journal of Financial Economics, 58, 81–112.

Claessens, S., Djankov, S., Fan, J. and Lang, L. (2002). Disentangling the incentive and

entrenchment effects of large shareholdings. Journal of Finance, 57, 2741-71.

Coase, R. H. (1937). The nature of the firm. Econometrica, 4(16), 386–405.

Cremers, K.J.M., Nair, V.B. (2005). Governance mechanisms and equity prices. Journal of

Finance, 60, 2859–2894.

Cummings, L. S. (2000). The Financial Performance of Ethical Investment Trusts: An

Australian Perspective. Journal of Business Ethics, 25(1), 79–92.

Demsetz, H. and Lehn, K. (1985). The structure of corporate ownership: causes and

consequences. Journal of Political Economy, 93, 1155–77.

Deng, X. , Kang, J.-K. , Low, B. , 2013. Corporate social responsibility and stakeholder value

maximization: evidence from mergers. Journal of Financial Economics, 110, 87–109.

Dharwadkar, R., George, G. and Brandes, P. (2000). Privatization in emerging economies: an

agency theory perspective. Academy of Management Review, 25, 650–69.

Diemont, D., Moore, K., & Soppe, A. (2016). The downside of being responsible: Corporate

social responsibility and tail risk. Journal of Business Ethics, 137(2), 213-229.

Dimson, E., Karakaş, O., & Li, X. (2015). Active ownership. The Review of Financial Studies,

28(12), 3225-3268.

Ding, S., Wu, Z., Li, Y., & Jia, C. (2010). Executive compensation, supervisory board, and

China’s governance reform: a legal approach perspective. Review of Quantitative Finance

and Accounting, 35(4), 445-471.

Doh, J. P., Littell, B., & Quigley, N. R. (2015). CSR and sustainability in emerging markets:

Societal, institutional, and organizational influences. Organizational Dynamics, 2(44),

112-120.

El Ghoul, S., Guedhami, O., & Kim, Y. (2017). Country-level institutions, firm value, and the

role of corporate social responsibility initiatives. Journal of International Business

Studies, 48(3), 360-385.

El Ghoul, S., Guedhami, O., Kwok, C. C., & Mishra, D. R. (2011). Does corporate social

responsibility affect the cost of capital?. Journal of Banking & Finance, 35(9), 2388-

2406.

Faccio, M., Lang, L. and Young, L. (2001). Dividends and expropriation. American Economic

Review, 91, 54-78.

Fama, E. F., & MacBeth, J. D. (1973). Risk, return, and equilibrium: Empirical tests. Journal

of Political Economy, 81(3), 607-636.

Ferrell, A., Liang, H., & Renneboog, L. (2016). Socially responsible firms. Journal of Financial

Economics, 122(3).

Filatotchev, I., Wright, M., Uhlenbruck, K., Tihanyi, L. and Hoskisson, R. E. (2003).

Governance, organizational capabilities, and restructuring in transition economies.

Journal of World Business, 38, 331–47.

64

Flammer, C. (2013). Corporate social responsibility and shareholder reaction: The

environmental awareness of investors. Academy of Management Journal, 56(3), 758-781.

Fombrun, C., & Shanley, M. (1990). What's in a name? Reputation building and corporate

strategy. Academy of Management Journal, 33(2), 233-258.

Friedman, M. (1970). The social responsibility of business is to increase its profits. New York

Times Magazine, 32(13), 122-126.

Gelb, D. S., & Strawser, J. A. (2001). Corporate social responsibility and financial disclosures:

An alternative explanation for increased disclosure. Journal of Business Ethics, 33(1), 1-

13.

Godfrey, P. C. (2005). The relationship between corporate philanthropy and shareholder

wealth: A risk management perspective. Academy of Management Review, 30, 777–798.

Godfrey, P. C., Merrill, C. B., & Hansen, J. M. (2009). The relationship between corporate

social responsibility and shareholder value: An empirical test of the risk management

hypothesis. Strategic Management Journal, 30, 425–445.

Goetzmann, W. N., Li, L., & Rouwenhorst, K. G. (2005). Long-term global market correlations.

Journal of Business, 78(1).

Gomes, A. (2000). Going public without governance: managerial reputation effects. Journal of

Finance, 55, 615–46.

Gompers, P., Ishii, J., Metrick, A. (2003). Corporate governance and equity prices. Quarterly

Journal of Economics, 118, 107-155.

Goss, A., & Roberts, G. S. (2011). The impact of corporate social responsibility on the cost of

bank loans. Journal of Banking & Finance, 35(7), 1794-1810.

Halbritter, G., & Dorfleitner, G. (2015). The wages of social responsibility—where are they?

A critical review of ESG investing. Review of Financial Economics, 26, 25-35.

Hemingway, C. A., & Maclagan, P. W. (2004). Managers' personal values as drivers of

corporate social responsibility. Journal of Business Ethics, 50(1), 33-44.

Hill, C. W., & Jones, T. M. (1992). Stakeholder‐agency theory. Journal of Management Studies,

29(2), 131-154.

Hong, H., Kubik, J. D., & Scheinkman, J. A. (2012). Financial constraints and corporate

goodness, National Bureau of Economic Research: NBER Working Paper 18476.

Hoskisson, R. E., Eden, L., Lau, C. M. and Wright, M. (2000). Strategy in emerging economies.

Academy of Management Journal, 43, 249–67.

Jensen, M. C., & Meckling, W. H. (1976). Theory of the firm: Managerial behavior, agency

costs and ownership structure. Journal of Financial Economics, 3(4), 305-360.

Jo, H., & Harjoto, M. (2014). Analyst coverage, corporate social responsibility, and firm risk.

Business Ethics: A European Review, 23(3), 272-292.

Johnson, S., Boone, P., Breach, A. and Friedman, E. (2000). Corporate governance in the Asian

financial crisis. Journal of Financial Economics, 58, 141–86.

Jones, G. R., & Hill, C. W. (1988). Transaction cost analysis of strategy‐structure choice.

Strategic Management Journal, 9(2), 159-172.

Jones, T. M. (1995). Instrumental stakeholder theory: A synthesis of ethics and economics.

Academy of Management Review, 20(2), 404-437.

65

Karna, A., Täube, F., & Sonderegger, P. (2013). Evolution of innovation networks across

geographical and organizational boundaries: A study of R&D subsidiaries in the

Bangalore IT cluster. European Management Review, 10(4), 211-226.

Khanna, T., & Palepu, K. G. (1997). Why focused strategies may be wrong for emerging

markets. Harvard Business Review, 75(4): 41–51.

Khanna, T., & Palepu, K. G. (2011). Winning in emerging markets: Spotting and responding to

institutional voids. World Financial Review, May–June, 18-20.

Kim, Y., Li, H., & Li, S. (2014). Corporate social responsibility and stock price crash risk.

Journal of Banking & Finance, 43, 1-13.

Kim, Y., Park, M. S., & Wier, B. (2012). Is earnings quality associated with corporate social

responsibility?. The Accounting Review, 87(3), 761-796.

Kitzmueller, M., & Shimshack, J. (2012). Economic perspectives on corporate social

responsibility. Journal of Economic Literature, 50(1), 51-84.

Krueger, Philipp, 2015, Corporate goodness and shareholder wealth, Journal of Financial

Economics, 115, 304–329.

La Porta, R., Lopez‐de‐Silanes, F., & Shleifer, A. (1999). Corporate ownership around the

world. The Journal of Finance, 54(2), 471-517.

Lacy, P., Hayward, R. (2011). A new era of sustainability in emerging markets? Insights from

a global CEO study by the United Nations Global Compact and Accenture, Corporate

Governance: The international journal of business in society, 11(4), 348-357.

Lee, D. D., & Faff, R. W. (2009). Corporate sustainability performance and idiosyncratic risk:

A global perspective. Financial Review, 44(2), 213-237.

Lee, M.T. (2016). Corporate social responsibility and stock price crash risk: Evidence from an

Asian emerging market. Managerial Finance, 42(10), 963-979.

Li, S., Fetscherin, M., Alon, I., Lattemann, C., & Yeh, K. (2010). Corporate social responsibility

in emerging markets. Management International Review, 50(5), 635-654.

Liang, H., & Renneboog, L. (2017). On the foundations of corporate social responsibility. The

Journal of Finance, 72(2), 853-910.

Lins, K. V., Servaes, H., & Tamayo, A. (2017). Social capital, trust, and firm performance: The

value of corporate social responsibility during the financial crisis. The Journal of Finance,

72(4), 1785-1824.

Luo, Y., & Tung, R. L. (2007). International expansion of emerging market enterprises: A

springboard perspective. Journal of International Business Studies, 38(4), 481-498.

Maignan, I., & Ralston, D. (2002). Corporate social responsibility in Europe and the U.S.:

Insights from businesses’ self-presentations. Journal of International Business Studies,

33(3), 497–514.

McCarthy, S., Oliver, B., & Song, S. (2017). Corporate social responsibility and CEO

confidence. Journal of Banking & Finance, 75, 280-291.

McGuire, J. B., Sundgren, A., & Schneeweis, T. (1988). Corporate social responsibility and

firm financial performance. Academy of Management Journal, 31(4), 854-872.

66

McKinsey Global Institute (2010). How the growth of emerging markets will strain global

finance. Retrieved on 15.09.2018 from https://www.mckinsey.com/featured-

insights/winning-in-emerging-markets/how-the-growth-of-emerging-markets-will-

strain-global-finance

McWilliams, A., & Siegel, D. (2001). Corporate social responsibility: A theory of the firm

perspective. Academy of Management Review, 26(1), 117-127.

Meyer, K. E., Estrin, S., Bhaumik, S. K., & Peng, M. W. (2009). Institutions, resources, and

entry strategies in emerging economies. Strategic Management Journal, 30(1), 61-80.

Morck, R., Wolfenzon, D., & Yeung, B. (2005). Corporate governance, economic

entrenchment, and growth. Journal of Economic Literature, 43(3), 655-720.

Morck, R., Yeung, B., & Yu, W. (2000). The information content of stock markets: why do

emerging markets have synchronous stock price movements?. Journal of Financial

Economics, 58(1-2), 215-260.

Mazodier, P., & Trognon, A. (1978). Heteroscedasticity and stratification in error components

models. Annales de l'INSE: Institut national de la statistique et des études économiques,

451-482.

North, D. C. (1990). Institutions, Institutional Change and Economic Performance. New York:

Norton.

North, D. C. (1994). Economic performance through time. American Economic Review, 84,

359–68.

Peng, M. W. (2003). Institutional transitions and strategic choices. Academy of Management

Review, 28, 275–96.

Peng, M. W. and Heath, P. (1996). The growth of the firm in planned economies in transition:

institutions, organizations, and strategic choice. Academy of Management Review, 21,

492–528.

Perez-Batres, L. A., Miller, V. V., & Pisani, M. J. (2010). CSR, sustainability and the meaning

of global reporting for Latin American corporations. Journal of Business Ethics, 91(2),

193-209.

Petrovits, C. M. (2006). Corporate-sponsored foundations and earnings management. Journal

of Accounting and Economics, 41(3), 335-362.

Pigou, A. (1920). The Economics of Welfare. London: McMillan & Co.

Prior, D., Surroca, J., & Tribó, J. A. (2008). Are socially responsible managers really ethical?

Exploring the relationship between earnings management and corporate social

responsibility. Corporate Governance: An International Review, 16(3), 160-177.

Qu, R. (2007). Effects of government regulations, market orientation and ownership structure

on corporate social responsibility in China: An empirical study. International Journal of

Management, 24(3), 582–591.

Russo, M. V., & Fouts, P. A. (1997). A resource-based perspective on corporate environmental

performance and profitability. Academy of Management Journal, 40(3), 534-559.

Sauer, D. (1997). The Impact of Social-Responsibility Screens on Investment Performance:

Evidence from the Domini Social Index and Domini Equity Mutual Fund. Review of

Financial Economics, 6, 137-149.

https://www.mckinsey.com/featured-insights/winning-in-emerging-markets/how-the-growth-of-emerging-markets-will-strain-global-finance



67

Shrivastava, P. (1995). Ecocentric management for a risk society. Academy of Management

Rrusseview, 20(1), 118-137.

Siegel, J. (2009). Is there a better commitment mechanism than cross-listings for emerging

economy firms? Evidence from Mexico. Journal of International Business Studies, 40(7),

1171-1191.

Thomson Reuters (2018). ESG scores methodology. Retrieved on 27.06.2018 from

https://financial.thomsonreuters.com/content/dam/openweb/documents/pdf/financial/esg

-scores-methodology.pdf

UN Principles for Responsible Investment (2017). About the PRI. Retrieved on 26.06.2018

from https://www.unpri.org/pri/about-the-pri

USSIF (2016). Trend Report Highlights. Retrieved on 26.06.2018 from

https://www.ussif.org/files/Trends/US%20SIF%202016%20Trends%20Overview.pdf

Utz, S. (2017). Over-investment or risk mitigation? Corporate social responsibility in Asia-

Pacific, Europe, Japan, and the United States, Review of Financial Economics (in press).

Waddock, S. A., & Graves, S. B. (1997). The corporate social performance-financial

performance link. Strategic Management Journal, 18, 303–319.

Welford, R. (2004). Corporate social responsibility in Europe and Asia: Critical elements and

best practice. Journal of Corporate Citizenship, 13, 31–47.

White, H. (1980). A heteroskedasticity-consistent covariance matrix estimator and a direct test

for heteroskedasticity. Econometrica, 817-838.

Williamson, O. E. (1975). Markets and hierarchies. New York: Free Press.

Williamson, O. E. (1985). The economic institutions of capitalism. New York: Free Press.

Young, M. N., Peng, M. W., Ahlstrom, D., Bruton, G. D., & Jiang, Y. (2008). Corporate

governance in emerging economies: A review of the principal–principal perspective.

Journal of Management Studies, 45(1), 196-220.

Yu, F. (2008). Analyst coverage and earnings management. Journal of Financial Economics,

88(2), 245–271.

Zhang, M., Xie, L., & Xu, H. (2016). Corporate philanthropy and stock price crash risk:

Evidence from china. Journal of Business Ethics, 139(3), 595-617.

https://financial.thomsonreuters.com/content/dam/openweb/documents/pdf/financial/esg-scores-methodology.pdf

https://financial.thomsonreuters.com/content/dam/openweb/documents/pdf/financial/esg-scores-methodology.pdf

https://www.unpri.org/pri/about-the-pri

https://www.ussif.org/files/Trends/US%20SIF%202016%20Trends%20Overview.pdf

68

Appendix

Appendix A: MSCI EM region clustering

Retrieved on 10.09.2018 from https://www.msci.com/market-classification

https://www.msci.com/market-classification

69

Appendix B: Thomson Reuters EIKON ESG score methodology

Retrieved on 14.10.2018 from

https://www.refinitiv.com/content/dam/gl/en/documents/methodology/esg-scores-

methodology.pdf

https://www.refinitiv.com/content/dam/gl/en/documents/methodology/esg-scores-methodology.pdf

https://www.refinitiv.com/content/dam/gl/en/documents/methodology/esg-scores-methodology.pdf

70

Appendix C: Variable overview

General

company

information

• RIC (Reuters Identifier code)

• Unique Company Name

• TRBC Business Sector Name

• Country

• Region: Asia, Americas, EMEA (see Appendix A)

Market &

Fundamental

data

All variables are downloaded in USD for comparability:

• 𝑇𝑂𝐵𝑄𝑖,𝑌 = 𝑀𝑎𝑟𝑘𝑒𝑡 𝐶𝑎𝑝𝑖𝑡𝑎𝑙𝑖𝑧𝑎𝑡𝑖𝑜𝑛𝑖,𝑌+𝐵𝑜𝑜𝑘 𝑉𝑎𝑙𝑢𝑒 𝑜𝑓 𝐴𝑠𝑠𝑒𝑡𝑠𝑖,𝑌−𝐵𝑜𝑜𝑘 𝑉𝑎𝑙𝑢𝑒 𝑜𝑓 𝐸𝑞𝑢𝑖𝑡𝑦𝑖,𝑌

𝐵𝑜𝑜𝑘 𝑉𝑎𝑙𝑢𝑒 𝑜𝑓 𝐴𝑠𝑠𝑒𝑡𝑠𝑖,𝑌

• 𝑅𝑂𝐴𝑖,𝑌 =Net income before extraordinary items𝑖,𝑌

Total Book Value of Assets𝑖,𝑌−1

• 𝑆𝐼𝑍𝐸𝑖,𝑌 = 𝑙𝑛 (𝐵𝑜𝑜𝑘 𝑉𝑎𝑙𝑢𝑒 𝑜𝑓 𝐴𝑠𝑠𝑒𝑡𝑠𝑖,𝑌)

• 𝐿𝐸𝑉𝑖,𝑌 =Total Reported Value of Debt𝑖,𝑌

Book Value of Assets𝑖,𝑌

• 𝐺𝐷𝑃𝑖,𝑌 = ln( respective GDP per capita in constant 2010 USD𝑖,𝑌)

• 𝐴𝐺𝐸𝑖,𝑌 = ln( 𝑌𝑖 − 𝑌𝑒𝑎𝑟 𝑜𝑓 𝑖𝑛𝑐𝑜𝑟𝑝𝑜𝑟𝑎𝑡𝑖𝑜𝑛𝑖)

• 𝐴𝑁𝐴𝑖,𝑌 = ln(𝑊𝑒𝑖𝑔ℎ𝑡𝑒𝑑 − 𝑎𝑣𝑒𝑟𝑎𝑔𝑒 𝑛𝑢𝑚𝑏𝑒𝑟 𝑜𝑓 𝑠𝑒𝑙𝑙 − 𝑠𝑖𝑑𝑒 𝑎𝑛𝑎𝑙𝑦𝑠𝑡𝑠𝑖,𝑌)

ESG data All raw scores are relative the respective firm’s industry peers and on a scale from 0-100. I

normalised them to range between 0-1.

• 𝐶𝑆𝑅𝑖,𝑌 =𝐸𝑛𝑣𝑖𝑟𝑜𝑛𝑚𝑒𝑛𝑡𝑎𝑙 𝑝𝑖𝑙𝑙𝑎𝑟 𝑠𝑐𝑜𝑟𝑒𝑖,𝑌+𝑆𝑜𝑐𝑖𝑎𝑙 𝑝𝑖𝑙𝑙𝑎𝑟 𝑠𝑐𝑜𝑟𝑒𝑖,𝑌

2

Environmental pillar categories:

• 𝑅𝑒𝑠𝑜𝑢𝑟𝑐𝑒 𝑈𝑠𝑒𝑖,𝑌, 𝐸𝑚𝑖𝑠𝑠𝑖𝑜𝑛𝑠𝑖,𝑌, 𝐸𝑛𝑣𝑖𝑟𝑜𝑛𝑚𝑒𝑛𝑡𝑎𝑙 𝐼𝑛𝑛𝑜𝑣𝑎𝑡𝑖𝑜𝑛𝑖,𝑌

Social pillar categories:

• 𝑊𝑜𝑟𝑘𝑓𝑜𝑟𝑐𝑒𝑖,𝑌, 𝐻𝑢𝑚𝑎𝑛 𝑅𝑖𝑔ℎ𝑡𝑠𝑖,𝑌, 𝐶𝑜𝑚𝑚𝑢𝑛𝑖𝑡𝑦𝑖,𝑌, 𝑃𝑟𝑜𝑑𝑢𝑐𝑡 𝑅𝑒𝑠𝑝𝑜𝑛𝑠𝑖𝑏𝑖𝑙𝑖𝑡𝑦𝑖,𝑌

Governance pillar categories:

• 𝑀𝑎𝑛𝑎𝑔𝑒𝑚𝑒𝑛𝑡𝑖,𝑌, 𝑆ℎ𝑎𝑟𝑒ℎ𝑜𝑙𝑑𝑒𝑟𝑠𝑖,𝑌, 𝐶𝑆𝑅 𝑠𝑡𝑟𝑎𝑡𝑒𝑔𝑦𝑖,𝑌

71

Appendix D: Breusch-Pagan test for Random Effects

Lagrange Multiplier Tests for Random Effects

Null hypotheses: No effects

Alternative hypotheses: Two-sided (Breusch-Pagan) and one-sided

(all others) alternatives

Test Hypothesis

Cross-section Time Both

Breusch-Pagan 3690.646 47.80646 3738.452

(0.0000) (0.0000) (0.0000)

Honda 60.75069 6.914222 47.84632

(0.0000) (0.0000) (0.0000)

King-Wu 60.75069 6.914222 12.84754

(0.0000) (0.0000) (0.0000)

Standardized Honda 60.90924 7.861114 35.20290

(0.0000) (0.0000)

(0.0000)

Standardized King-Wu 60.90924 7.861114 10.31710

(0.0000) (0.0000) (0.0000)

Gourierioux, et al.* -- -- 3738.452

(< 0.01)

*Mixed chi-square asymptotic critical values:

1% 7.289

5% 4.321

10% 2.952

72

Appendix E: Heteroskedasticity and Normality in the fixed effects OLS model

Dependent Variable: RESID^2

Method: Panel Least Squares

Date: 09/20/18 Time: 12:55

Sample: 2010 2016

Periods included: 7

Cross-sections included: 657

Total panel (unbalanced) observations: 3800

Variable Coefficient Std. Error t-Statistic Prob.

C 1.135483 0.085856 13.22538 0.0000

CSR^2 0.041158 0.033243 1.238071 0.2158

ROA^2 3.472582 0.264709 13.11850 0.0000

LEV^2 -0.271950 0.072735 -3.738925 0.0002

SIZE^2 -0.001546 0.000137 -11.24631 0.0000

GDP^2 -0.003426 0.000466 -7.349656 0.0000

AGE^2 0.003248 0.001799 1.805608 0.0711

R-squared 0.118056 Mean dependent var 0.132740

Adjusted R-squared 0.116661 S.D. dependent var 0.460751

S.E. of regression 0.433042 Akaike info criterion 1.165875

Sum squared resid 711.2829 Schwarz criterion 1.177375

Log likelihood -2208.163 Hannan-Quinn criter. 1.169962

F-statistic 84.62148 Durbin-Watson stat 1.240603

Prob(F-statistic) 0.000000

73

Appendix F: Heteroskedasticity and Normality in the WLS model

Dependent Variable: RESID^2

Method: Panel Least Squares

Date: 09/20/18 Time: 12:46

Sample: 2010 2016

Periods included: 7

Cross-sections included: 657

Total panel (unbalanced) observations: 3800

Variable Coefficient Std. Error t-Statistic Prob.

C 3.80E+41 1.77E+41 2.144670 0.0320

CSR^2 -5.23E+40 6.87E+40 -0.761587 0.4464

ROA^2 1.57E+41 5.47E+41 0.286719 0.7743

LEV^2 -1.67E+40 1.50E+41 -0.111184 0.9115

SIZE^2 -5.03E+38 2.84E+38 -1.769385 0.0769

GDP^2 -1.36E+39 9.63E+38 -1.416655 0.1567

AGE^2 1.52E+39 3.72E+39 0.408673 0.6828

R-squared 0.001958 Mean dependent var 1.94E+40

Adjusted R-squared 0.000379 S.D. dependent var 8.95E+41

S.E. of regression 8.95E+41 Akaike info criterion 196.0343

Sum squared resid 3.04E+87 Schwarz criterion 196.0458

Log likelihood -372458.3 Hannan-Quinn criter. 196.0384

F-statistic 1.240132 Durbin-Watson stat 1.409558

Prob(F-statistic) 0.282280

Date post:	09-Sep-2020
Category:	Documents
Upload:	others
View:	2 times
Download:	0 times

Master Thesis - RUN: Página principal · 2019. 6. 26. · Master Thesis The effect of CSR...

Documents